- | º º º - º º º º º ſº º º º º º - - . - - º º º - - - - - º º º º º º º º: º º º º … º º º º º º º º: º º º --- º º º | - º º º: º ------- - º º º º º º -º-º-º: º º º --- - º º º º º º º º º º º º ------- º º º º º º º º: º: º º º º º º º º º º º º º º º º º º º: sº º: º º º º - ºº: º º ºº º º: º - º º º º º º: º º º º º º º º: º º º º º º º º º º ºrs: º º º º … º: º ºº: º º º: sº º º º º Ea. º º º: - º - º º º º º º º º --- º: º º º º º º º º º º º ºº: º: º - º: º - --~~~ º º º º - º º º º: - - º º º, º º º: º: º º º --> º º º º º º º º º º º º º º º º º º - º º º º º º º ºº:: ºº:: º º º --- º º ºº: - º º º º: - º, |-- º º: ºº º º: º º º - --- º º º º º: º - . -- - ºº: º º: º: º: - º º º º º º º T º º - º - º º º º -- - º ºº i. º º º º º º º º º --- º º - º º º º: - º º º º º: º º º º º º º - ºn º º º º º º: º - º: i- º º º º º-ºº: T º º º º º º re º - º º: º: º º º º º º º º º º ºº:: - - º º º º º Eºsi. º º º: -----~~~~~ †: º º *** º T º: T º º: º: º º: º: º: º º º: º º º --- º º T º: º º - --- º: º º º º º: º º: º º º º º: º ºº: º - º, º - º º º º º: º ------ - - ** ** º º: ºº:: T º º º - ºº: º º º: º º - ºº:: º º º: º º º º º: - º º - º º - º Tº -- º, º º º º º - | -- in º º - º º * rºº º ****** º º º º: - º º: ºº: º º ºº:: **** º - -T- *** º º - º - º ºn. º º º ºº:: Tºº- º º: -- º º º º º º º: º: º: -- º º º º º º º - ºr. º º º º º - ----- ------e. º º º º -- - * ** --- º - º: - * **** - - º - ºº: … " ", - - - --- º - - - º - º --- º º º - - º º - - - - - º * ºil. --~~~ º - ºº:::::::: º --- - - - * - **- º º º º: º º º º:ºº º T º ºº: º º º: ºil. - *- º º T |Tºº º º - cº º - º º **º-ºº: ..º.º. º º º º - º º º ºº. º ºr. Tºº-º-º: - º: - ºº:: --- º º º | º - º - - º º º -- º **** - º º -- ** --- º º - º ******** º ºr. º º º ºr. * - º … º --- ºº:: - ºr rºad. * - - - - - - Cº-ºº- º º- º º: - - - TS ( > s é & fº - | 000s 8 USES OF PERIODATES AND PERIODIC ACID IN ANALYSIS . * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * A DISSERTATION sk + k + k + k + k >k k >k + k + k + + Presented in the Partial Fulfillments of the Requirements for the Degree of Doctor of Philosophy in the University of Michigan. By LUCIEN H. GREATHOUSE *k # * * *k sk ºk ºk ºk ºkºk k ≤ x ≤k × x >k k >k k + k 1917. × Skºk + k ≤k sk The author wishes to thank first, Dr. H. H. Willard, under whose directions these investigations were made, for his assistance and suggestions, and also many other members of the chemical faculty, especially Professor W. G. Smeaton and Mr. R. K. MacAlpine, for their kindly advice. CONTENTS. Int; roduct ion * - e. e-4 e º e - e < * * Historical e-º-º-º-º-º-º º tº e - e-4 Theoretical. * -- e-º-º-º-º-º-º-º-º-º: Experimental Precipitation Methods Qualitative Examinations Iron . . . . . . . . . . Potassium. . . . . . . . Oxidation Methods, General Reactions, . . . Standard Solutions . . . Chromium and Wanadium . Iodine , . . . . . . . Bromine . . . . . . . . Chlorine. . . . . . . . . º e Rate of Reaction of Periodic and iodine . . Summary . . . . . . . . . . . . . References º º * -- - - --> º -e - * 33 24 39 Full Page Curves and Illustrations. Following page º Curves of Electrometric Titrations of Wānadyl Sulfate . . . . . . . . . . . 61 Curves of Reaction of K2H3106 and KI . . . . . 65 Apparatus for reaction of iodine and - periodic acid -º-º-º-º-º-º-º-º-º-º: 79 Curve of molecular condictivity of periodic acid at 45°. . . . . . . . . 88 USES OF PERIODATES AND PERIODIC ACID IN ANALYSIS. INTRODUCTION. It is nearly eighty five years since a series of in- organic compounds has been described that contained as an acid constituent some radical derived from the heptoxide of iodine. During that time their properties have been studied extensively. These are unique in many regards. Chemically, compounds of this series are active and their relationships extremely varied. Yet to the present day, there is hardly a known application in technical or scientific procedure of the effects to be obtained with these substancess As the author, among others, has found, certain types of their reactions are definitely quantitative, and readily subject to control by the chemical manipulator. Such are in particular those in which the periodate functions as an oxidizing agent. On the other hand, in the simple metatheti- cal reactions with various bases or salts, the periodates present a maze of complication, only comparable in inorganic chemistry, to that found among the natural silicates. It is perhaps this condition and the difficulties encountered in the experimental study of it that have kept the entire field so far neglected. For the greater part of the investigation of the periodates has been concerned with determinations of the intricate stoichometric ratios of the salts and the basic- ity of the free acid. Much of this has yielded indefinite and 3. unsatisfactory results. HISTORICAL • In considering the previous literature chronolºgical order will not be strictly observed, but division will be made into investigations concerning reactions wherein the iodine remained heptavalent, into those involving oxidation processes, and into such few as have related to analysis. F. Ammermüller and G. Mangus ; are to be credited with the original work which brought to light the existence of the periodates. The analagous series of the perchlorates was already known. These chemists had been preparing sodium iodate by oxidizing iodine in aqueous solution with free chlorine , with addition of sodium carbonate as the reaction proceeded, to neutralize the ticids formed. When however they had used initially a large excess of alkali, the reaction proceeded dif- ferently and the sodium salt of a new oxy-acid of iodine was precipitated as a heavy white crystalline powder. From a solution of this substance in nitric acid they at once pre- pared, by adding silver nitrate and evaporating, the simplest silver salt. This is now named the metaperiodate, Agioa. They ignited a sample of this compound, from the weight of the residue of AgI, and the volume of the oxygen evolved, es- tablished its formula, and demonstrated the analogy of these salts to the perchlorates. But further, these authors recognized the existence of a type of salt containing two equivalents of the base to one atom of iodine • They determined that the sodium compound first prepared was of this sort. In addition they described the potassium salt and two hydrates of the silver salt constitut- ed in this same proportion. These silver compounds were ob- tained by boiling AgIoa, by which all the silver was thrown down as the dibasis salt leaving one half of the periodic acid of the original compound in solution. In this way Ammermüller and Magnus also prepared the free acid. A few years later, Benesser.)" Working under Liebig, undertook further study of the periodates. His first step was to prepare the free acid by decomposition of lead period- ate with a small excess of sulfuric acids Theoretically this was an improvement as regards yield, over the method used by Ammermöller and Magnus. However he found at once the diffi- culty of crystallizing the free acid from solutions containing any appreciable amount of sulfuric acids Bengleser gave the melting point of the free acid as 130°C, approximately that agreed upon by later investigators. But in stating the temp- eratures for dehydration and decomposition, he opened a dis- cussion which has hardly been completely settled to the pres- ent time. His figures for these values were considerably above any accepted today. Bengieser examined the oxidation reactions of periodic acid with a number of substances. His results will be considered later. Bengieser also prepared the mercurous and mercurio per- iodates, besides the lead salt and showed that these three, like the dibasic silver salt precipitated from cold solution 4. by Ammermüller and Magnus, all lost water in boiling solu- tion and darkened in color. In 1838, Rammelsburg ; showed that by gentle ignition of the iodates of the alkaline earths, the iodine and oxygen were volatilized in part, but that con- siderable amounts of highly basic periodates were present in the residue. He examined this process rigidly to deter- mine that this reaction actually occurred during the ignition, and that the periodate formation did not take place by some subsequent reaction as the ignition residue was being brought into the solution, - After these investigations, following close on the orig- inal discovery of the series, the chemistry of this field re- º mained largely untouched for thirty years. The only notable --- 4. exception is the work of the Frenchman, Langlois, ) published in 1852, This author examined thoroughly the known methods of preparing the sodium salt and the free acid. He stated the ëssential conditions for the oxidation of iodate to periodate in alkaline solution with free chlorine, for maximum yield, to be that concentration of caustic be high, and solution near boilings For preparation of the free acid, Langlois peeferred the method of Ammermºiller and Magnus using the silver salt. He found the decomposition of the lead salt With sulfuric acid as proposed by Bengieser, unsatisfactory, in that the product obtained was impure. Langlois prepared and analyzed for the first time, the monobasic ammonium salt, the dibasic barium, calcium, stron- tium, and magnesium salts, the tribasic lead salts, and the tetrabasic zinc and copper salts. He stated that ferrous and 5. manganous salts were oxidized by periodic acid. He also prepared periodates of a number of the alka- loids, in particular that of quinine, which he analyzed and to which he ascribed a formula. A book on crystallography published in 1857 by Rammels- burg and Marignac ; described the crystal form of a few periodates. In 1863, Kammerer ; gave periodic acid as the product of the reaction of perchloric acid on iodine. This statement; was later discredited by Michael and Conn (1901). * - The year 1867 marks the beginning of a decade of vigor- ous activity in attempts to solve the complexities of this fields By far the largest part of this work was directed towards determining, in one way or another, the 'normal' basicity of periodic acids With the aid of this it was hoped there might be evolved a classification for the many periodates previously known or discovered in the course of these investigations. In the work of this period the one name which stands - 9, 10, 10a preeminent is that of C. Rammelsburg. ) Two other authors, 8 - 7 C. G. Lautsch ) and J. W. Fernlunds ) also communicated in 1867 results of investigations along very similar lines. For our purposes it is not necessary that we follow in detail the elaborate studies of Rammelsburg, nor those of these contemporaries. Their whole method was to prepare salts of periodic acid, from solutions of various concentrations of free acid or alkali and of the base in question, and from their analyses to determine formulae. In the course of a few years 6. they prepared several such compounds with each of the common bases with which periodic acid enters into the stable combina- tion. - - 10 In the table by which Rammelsburg ) finally sums up and systematizes the compounds Whose existence he regarded as established in these or earlier researches, there are ten classes of salts. These include some forty five different periodates of the alkalis, alkaline eatths, silver, mercury, copper, lead, zinc, cadmium, nickel and iron. He named the several classes, first normal, being of the KIO4 type, and thereafter according to the fraction of an atom of iodine combin- êd. With one equivalent of base, three-fifths, four-sevenths-, half-, two-fifths—, three eighths-, third-, fourth-, fifth-, and sixth-, periodates. It is evident that, especially for bi- and trivalent cations, some of these ratios must be used to indicate compounds of very high formula weight. In addition some of the most complicated forms carry from fifteen to as many as fifty molecules of Water of hydration. Rammelsburg regarded at this time that his results did not justify any assumption other than that of the Rio, type was the normal form and the others basic modifications. In general only the less basic of the salts were well- defined crystalline bodies. The highly basic combinations especially of the heavy metals were only obtained as fine powd- ers or else as gelatinous precipitates. 3 The basic salts. With the exceptions of the potassium and ammonium compounds Were uniformly insoluble. The normal periodates were of course obtained by evaporation of acid solutions, the basic, as a rule, from neutral or alkaline solutions. Lautsch, Fernlunds, and Rammelsburg obtain contradiot- ory results in regard to the silver salts, which on this ac- count were examined at great length. Fernlund's entire paper dealt with these, and Rammelsburg jº a little later communicated separately on them in answer to the work of the other two, Whatever the reason may be, the composition of these salts present the utmost variety and uncertainty. At the end of his paper on the silver salts, Rammels- burg refutes another statement made by Lautsch, that free H5106 loses water in the air, and then only liberates oxygen 200° to 210°C. Rammelsburg found that Hsios is stable in air, but decomposes near its melting point, 130°C. In still another paper, Rammelsburg 5*.svers. to the subject of the ignition products of the periodates. Several authors writing within a few years after Rammelsburg's work, without adding much experimental data, set about to interpret his work differently. K. Ihre jº repeated a number of the prep- arations and analyses described by Rammelsburg, and in his work added another touch to the silver salt discussion. He believed that periodic acid could have a basicity of one to five, o, Bicºlaj "sonºurse in these views. H. Kammerer jº, also preparing several of the basic periodates, took the view directly opposed to Rammelsburg's, that the five-basic salts constituted the normal type. He proceeded to reformulate according to this scheme many salts which Rammelsburg had con- sidered hydrates; for example, Ba21209 - 3 H20 became BaháI06. 16 The work of J. Thomsen, ) in 1873 and 1874 is related, by its place in the literature, to that of the other authors just considered, but marks a transition to a more modern period in the method of attack on the problem of the basicity of periodic acids Thomsen measured several physical values, in particular the heat liberated in neutralizing with succes- sive equivalents of base. He concluded that periodic acid is normally dibasic. He also regarded that periodic acid had twice the molecular size previously ascribed to it, using the formula. H T209-3H29. In this he came into discussion With essarº who held to the older view. In 1885, onwalaº began the investigations of the basic- ity of periodic acid by the methods of modern physical chemistry. He determined the conductivity and dissociation with varying dilutions and found that H5106 is polybasic. P. raise.” using the same type of method concluded that the normal period- ates are the monobasic salts. Further, other authors began anew to study the prepara- tion and composition of the polycasio salts. C. W. zºº)" in two papers described new compounds with potassium, silver, lead, ferrous and ferric iron, copper, nickel and cadmium. This author used the nomenclature of Blomstrand, whereby Ram— melsburg's normal, half-, third-, fourth-, and fifth- periodates become respectively meta-, diortho-, meso-, dimeso-, and ortho- periodates. He contested the view of Thomsen that the dibasic salts represent the limiting condition for normal salt. He even stated that compounds exist of the same composition, in 9. effect inorganic isomers, which exhibit a marked difference in the ease with which water may be driven off, according as the hydrogen atoms are in the water of hydration or united as hydroxyl to the iodine, as Ag+I209, H20 and Agafiro5. Raº about this tiie described a mercuri am— - 24, 34a, 39 monium periodate, 3NHg2. IO4. NHg20H. NH4OH.3H2O. Blomstrand ) 1886, 1889, 1892, communicated three papers in this period on the same general subject. He believed with Kimmins that the periodates were salts of heptavalent iodine, with all five hydrogens replaceable. In the last paper appearing in 1892, he presented extensive data in support of this view from study of a series of complex periodo-molybdates. The compounds of this sort formed most readily he found to be of the type formula. Rºos-smoog. He prepared the simple salts of this ratio of the alkalis and the alkaline earths; also a number of mixed salts of two bases, These were in general hydrated with nine to seventeen molecules of water per acid radical. By treatment of salts of this type with ammonium he was able to remove molybdic acid, and obtained salts With four molecules and with one molecule of MoC, 3 Closely related to this last work were the investiga- tions of Rogenheim and assº.” some years later. These author's first reexamined the threadbare subject of the silver salts. They showed that Ag5103 W3.3 the end-product of the treatment of any of the less basic silver periodates with an excess of boiling water, ammonia or silver nitrate. Secondly, they prepared a series of periodotungstates, corresponding close- ly with that of the periodomolydates. They found a number of 10. t; he pentabasic hexatungstate derivatives and a few of the monotungstate series. These authors also utilized physico-chemical methods to decide the question of the basicity, determining conductiv- ity and effects of electrolysis with Na5106.6Moog. Their con- clusions are that periodło acid is normally pentabasic, but with strong bases it forms a series of monobasic stable normal salts, the metaperiodates. One more investigation of purely analytical type remains to be considered. In 1903, gaolitti ń published on the poly- basic periodates of lead and copper. He founăºhe copper salts, as Rosenheim and Liebknecht for the silver, that excess of am- monia or of a suspension of cupric hydroxide gives a pentabasic compound. But trying similarly hydrolysis of the lead salts, he obtained only tribasic compounds. A number of attempts to decide this question have been made in recent years, using purely physico-chemical methods. After the work of Ostwald and Walden already mentioned, there appears among these first a note by J. M. Crofts. jº Using the method of Lowenherz of measuring the depression of the freezing point of Glauber's salt he found that KIOA was in 46 monomolecular aggregation in that solution. Anstruc and Murco ) studied the reactions of different indicators on neutralizing periodic acid with solutions of various alkali and alkaline earth hydroxides. They obtained sharp end-points with all hydroxides and methyl orange on addition of one equivalent of base. But with phenolphthalein the end-points appear at ll. | different stages from 1.5 to 1.7 equivalents of base added when the alkali hydroxides are used, and from 1.9 to 2, 3 equivalents when alkaline earth hydroxides were used. Lamb * Hie, by careful drying under reduced pressure, but was unable to stop the dehydration from Hs”g, at any stable condition corresponding to H3105. Giolitti ) in a separate article from that previously considered, tried the use of electrometric titration of periodic acid with alkali, measuring potentials against the hydrogen electrode. He found it necessary to use a smooth platinum electrode, since periodic acid was reduced rapidly by the usual Hydrogen-platinum black electrode. 68 - In 1913 Eugene Corneo ) determined the freezing points, also the indices of refractions of periodic acid solutions successively as they were neutralized by various bases. As the potassium and ammonium salts were insoluble in acid solu- tion, and sodium salt in alkaline, it was impossible to carry through a complete series with any one base. The results When plotted, showed sharp breaks by both methods with sodium hydroxide where one equivalent of base had been added, and indicated fairly by freezing point method, with potassium hy- droxide, a change on addition of two equivalents. A year later Rene' bºrº examined the dissociation of periodic acid by a new method. He carried out a simple titration, but indicated the successive end-points, as the several hydrogen atoms were neutralized, by the appearance of a few cubic centimeters of insoluble oil in contact with the solution. Any increase in the alkalinity of the solution 13. - decreased the surface tension of the oil-water surface, and in turn increased in noticeable fashion the state of division of the oil. This method indicated that even the third hydro- gen atom dissociated somewhat. Occasional references to the oxidation reactions of per- iodates appear very early in the literature, but the bulk of the work of these is of quite recent date. Ammermüller and Magnus mentioned the reaction with hydrochloric acid to give chlorine and iodine chloride, Bengie ser stated that aqueous periodic acid oxidized elemental phosphorus, and a number of metals, sulfur, also tartaric, oxalic, formic and acetic acids. It appears that he was in error as regards the last two. Lang- lois and Rammelsburg both found that periodic acid was reduced by 'strong" alcehol, and the former added ether. Observations in this regard will appear in another place. Rammelsburg could not obtain ferrous, manganous or cobaltous periodates because of oxidation reactions and disputed the claim by Lautsch of a cobaltous salt, In 1887, sailºr, ſº presented at length a study of the rate of reaction of periodic acid and sulfur dioxide. Grätzner (1896) jº found that periodic acid was reduced only to iodic acid by formaldehyde. Some years later Brönner and ºne.” contradicted this statement with the claim that iodine was the reduction product , An explanation of this mat- ter is that Brºnmer and Mellet had in their solutions nitric acid in high concentration which itself was reduced by the formaldehyde to nitrious acids This last substance does reduce 13. | - both periodic and iodic acids to iodine, These same authors mention reducing periodate with hy- º droxylamin and hydrazin salts as also later Browne and Sheterly. jº 45, 52, 53a, 53, 54a, About 1900, Erich Müller) began his extensive investigations of the electrochemistry of the iodates and periodates, The chief purpose of this work was to devise successful electro- lytic oxidations for the production of these salts and of free periodic acid. But it also threw valuable sidelight on the oxidation potentials involved, and particularly in important interactions of compounds of iodine in different states of oxidation. 13 - J. Phillip ) had much earlier obtained the general re- action 3KIOAt KI = 4KIO3. E. Pechard jº had decided that starting with the metaperiodate and iodide, without other ad- dition of base or alkali, the reaction proceeded rapidly at first and was then retarded by the base liberated. He pro- posed an exact formulation for this process, with which Gar- assºciº-mºniº disagreed. - walls.” encountering the same reaftion in connection With his determination of the oxidation potentials of the series in normal KOH, concluded that to support his thereº- cal conclusions, it should occur in alkaline solution. Yet in fact it did not progress to a measurable degree under these conditions. He then set about to find catalysers which would increase the speed of the reaction until equilibrium was ap- proached in finite time. He did obtain this effect by dimin- ishing the alkalinity of the solution, by heat, possibly light 14. and by letting it accur in the presence of platinum, or bet- ter platinum black. W. *** discussed this reaction further, following up the controversy between Pechard and Garza rolli-Thurnlackh. He believed that both had neglected the effect of CO2 from the air on the progess of the reaction and could not confirm exact- ly the results of either. Auger treated NaIO4 solution of a known content with somewhat more than enough KI to react With all of the periodate oxygens He then titrated aliquot portions of this solution at successive times, directly with thiosulfate, and found the reducing reagent used to correspond, immediately on mixing, to the reaction 3NaIO4+ 3NaI - 2Nashg1064 NaIO3 + 3I. Pechard had formulated the reaction similarly, on the basis of results obtained by the same titration but stated that the re- action was not completed within an hour. From experience gained in this work it has not been found possible to interpret the meaning of results obtained by the titration as described. A few other references to oxidations by periodates appear in the literatures The difficulties met with by Giolitti in use of the platinized hydrogen electrode have already been mentioned. T. Tanatar ) described the reduction of periodic acid with 58 H2O2. o. Loew and K. Aao ) in connection with a study of the catalytic action of platinum black stated that KIO3 was reduced by glucose in the presence of this substance, while no catalyzer was necessary for the reaction with KIO4. D. Vitali ) examined the reactions of periodates with a number of fairly active reducing agents. The literature of the determination of periodates, apart 15. from general methods for iodine in its oxy-derivatives, and **i. of their use in analytical chemistry, is not large. Kammerer ) proposed to utilize the fact that barium iodate was transposed by amonium carbonate while the periodate was not, to effect a separation of the two radicals. T. Fairley (1876) suggested the precipitation of the insoluble sodium salt with iodide and excess of strongly alkaline hypochlorite, as a qualitative test for that element. E. Pechard jºined two possibilities for the determination of periodates. One was by the oxidation of standard oxalic acid and titration of the excess. The reaction required the addition of MnSO4 as a catalyzer. The other method was the titration of the first hydrogen atom in periodic acid With methyl orange as an indicator. The work of Anstruc and Murco (1902) already mentioned, indicated the same procedure. Beneas.” used potassium periodate as a test for Mn and Nis The one existing method of real merit for determination of the perio date radical in presence of the iodate is that wherein the iodine liberated in neutral solution with an excess of iodide, is titrated with standard reducing agent, either thiosulfate or arsenities The principle was first stated by Pechard ye in 1899, However, Mtiller and Fºssesses." in 1903 presented the method in detail. The most important point in the procedure is the control of the hydrogen ion concentration during the titration. If the solution is at all acid, the iodate will begin to react with the excess iodide, if alkaline the appear- ance of the oxidizing power of the periodate as free iodine will not be complete. Möller and Friedberger directed the neutralization with sodium bicarbonate, but this reagent may 16. º leave either free carbonic acid or normal carbonate, depending on the condition of the original solution, hence is not en- tirely satisfactory • These authors also tried sodium acetate- acetic acid mixtures but did not find this better. A decade later Mºller and Jacob jº tested the effect of adding an excess of standard arsenite directly to strongly alkaline solution, then neutralizing and titrating with standard iodine. This, too, gave uncertain results. But within the same year, Möller and Wegelin ;" described the use of borax and boric acid as a *buffer" combination, The most definite advantage of this pro- cedure is that the free boric acid can be added with impunity, since even in saturated solution it will not liberate iodine from an iodide—iodate mixtures We have used this method continually in work described herein, and found it a particularly satisfac- tory determination. l?. - THEORETICAL . It is evident that according as there are two general classes of reactions by the periodates, there will be two types of determ- imations which may be based on these reactions. Firstly, there will be methods, in which the constituent to be estimated will be precipitated from the solution as an insoluble compound. As the insoluble periodates are numerous, such possibil- it is s are many. If the precipitate is of constant composition either it may be weighed or the oxidizing power of the combined periodate may be determined by titration, if not, only a separat- ion can be devised. Secondly, there will be methods, in which some oxidizable sub- stances will react with a definite amount of standard periodate solution. Inasmuch as there can be no purpose in using a periodate to do work that an iodate would do as well, only titrations will be of consequence in which the reacting periodate is reduced to iodate. To perfect the latter type of method, it will be necessary to find oxidation reactions with periodates that go to completion readily ; and which have satisfactory end-points and to prepare stable standard solutions of periodate. It would appear at first , that because of these unsolved problems for the second type, the development of methods of the first type would offer greater promise. We shall find, however, that there are practical difficulties in the handling of these periodate precipitates, that will at least equalize the prospect of general usefulness of the two types of methods. But before any methods at all may be considered, it will be necessary to know what periodate compounds by virtue of their ease of preparation, as well as their chemical properties, Will be available for reagents. 18. Reagent Forms For the purposes of reagents it will be necessary to use compounds which will give moderately concentrated aqueous solu- tions. Here we are at once limited to the salts of the alkalis and the free acid. Potassium Periodates. The potassium metaperiodate is the only salt that can be obtained from the supply houses. It is prepared by the oxidation of alkaline iodate solution, by electrolysis as described by Muller or by the older method of passing chlorine gas into the solution. In either case the salt KIO4 is precipitated quite completely on acidifying. The high temperature coefficient of solubility renders purification by recrystallizing, easy and effective. Analysis of a sample obtained from Kahlbaum, used extensively in this work, was as follows :- Theoretical I II. III Average * º º % Loss in wt. at llo . OOO . O26 . O35 . O33 . O3 degrees % K2O 20.48 39.37 30.30 -- 30.24 % I2O7 79.53 79.5l 79.51 79.50 79.5i Total 100. OO - 99.78 The apparent low proportion of base probably arises from a small impurity of sodium, which was detected qualitatively. This, how- ever, was of no consequence in regard to it 3 use as a reagent. The salt dissolved to a permanently clear solution, which itself indicates practical freedom from bases other than the alkalis. The iodine hept oxide content given here is determined from t it ration with ståndardi Na2S293, of the total iodine liberated in acid solution with an excess of KI, jº, corresponding to the ºn 19. total oxygen present, or eight equivalents per molecule of KIO4. No iodate could be detected qualitatively in the fresh solution of the salt. Furthermore, it was repeatedly demonstrated, later, in making up standard solutions for reactions involving only the periodate oxygen, that this corresponded exactly to two equivalents per molecule of KIO4 or one quarter of the total oxidizing power. It is then evident that as a net result, such a preparation of K104 may be used as an exact standard by weight either for ordin- ary iodimetry, or for the special titrations, using reduction to the iodate only, with which we shall deal largely. The actual data. and the description of the methods, by which the fundamental stand- ards for the above volumetric determinations were established, as also details of the work in perfecting a stable and practical stand- ard solution of potassium periodate will be given in another place. In one type of standard solution the dibasic salt, potassium dimetaperiodate K2H3106 was used. This salt is quite soluble at room temperature, and forms a stable solution in the presence of a slight excess of alkali. KIO4 is soluble to less than fortieth mol- ar concentration at 35°C, and the metaperiodates in Water solution decompose steadily to iodate and ozone. Sodium periodates. The preparation of the insoluble basic sodium periodate has already been outlined from the literature • The method has also been studied fully in this labaratory for the preparation of the reagent for the colorimetric determination of manganese. The salt precipitated from alkaline iodate solution by chlorine has a com— position between Nagł2106 and Na2H3106, depending on the concentra- tion of sodium hydroxide. The iodine is removed from solution quantitatively when the alkali concentration is high. This insolu- bility of the basic sodium salt renders the yield on 2O. this method of preparation excellent, but makes it necessary to convert to the soluble metaperiodate for most purposes. This is done by dissolving the basic salt in a slight excess of nitric acid and evaporating on the steam bath. In one preparation of this sort 84% of the original periodate was obtained in the first crop of NaIO4. This material analyzed for ico, showed :- - I - II Theoretical for NaIO4. % of Igor 35. 6l. 85.61 85 - 59 This salt is stable in air. Periodic Acid. The free acid was used largely as a reagent in precipitation methods. It was, however, not to be obtained on the market. Two general methods for the preparation are recorded in the literature, first, decomposition of the barium salt with sulfuric acid as the silver salt with bromine, second, the electrolytic oxidation of iodic acid using a lead peroxid anode, porous cup diaphragm and sulfuric acid as the cathode chamber ***** This method was described by Muller and Friedbergerjin consider- able detail. Later ºullºuated the theory of the oxidation in presence of the PbO2, and concluded that a chemical action rather than a greater overvoltage makes this electrode effective, when a platinum anode practically failed entirely to give this oxidation. However, this procedures has one serious defect in that some sulfuric acid always passes into the anode chamber, and its presence interfere's in the crystallization of periodic acid. This difficulty was overcome in two different Way 3. In the earlier preparation3, the anode liquor was diluted and treated with a calculated amount of barium periodate of known º 31. - -- - barium content. Thus the two general methods were combined. This grocedtire in all was lengthy. The product was reasonably pure, though a small contamination of barium was always present. This Was not serious for work in acid solution. However, for the deter- mination of potassium where it was desired to evaporate the filtrates and determine sodium and traces of potassium as sulfate residue, it was necessary to prepare a purer product. For this purpose a modification of the original electrolytic method, devised by Dr. H. H. Willard, served excellently. Instead of a Jead peroxide coating carried on metallic lead being used as an anode, a platinum gauze was coated with lead peroxid. This type of electrode had already been used by Muller for another purposes . In this case, it had the important property of remaining unattacked by nitric acid. Hence the latter acid * could be substituted for sulfuric, and gave a mother liquor 5 much better even than pure water for the separation of periodic acid. Aqueous periodic solutions on evaporation assume a syrupy consistency before they are sufficiently concentrated for effecient crystallization. From concentrated nitric acid on the other hand, the separation is quite sharp since the mother liquor has but slightly greater viscosity than pure nitric acid. The best procedures for this crystallization is to concentrate in an open dish as rapidly as possible at 100 – 110 degrees C until crystallization almost begins, then to cover the dish and allow the solution to cool to room temperature without stirring . A slight amount of decomposition to iodic acid will occur in any case, but apparently this is a rather slow process and best 33. kept to a minimum by rapid evaporation. The periodic acid obtained by crystallization is entirely free from iodic. This crystallized acid was removed by centrifuging in platinum Cl193 - The crystals were washed with a little concentrated nitric acid. The acid was freed from adherent traces of nitric acid by drying under pressure of about 1 mm, with freshly ignited CaO. The resultant material gave the theoretical value for I 207 content, when calculated as H5I06. The test used for the detection of iodate in the periodate prepared is that first used by Langlois for the same CullºC 36 . It is based on the difference in solubility of the silver - salts in dilute HNO3. This is best carried out in the following manner. One half gram of the periodate to be tested is dissolved in 35 cc of water. If the free acid is being examined, this of itself will suffice to keep the silver periodate dissolved in warm solution; if a salt, an amount of nitric acid, usually 3 - 4 drops, equivalent to the base present, should be added. Then 3 - 3 drops of } silver nitrate are adidied and the solution warmed until the yellowish brown silver periodate reaissolves, which should occur readily at 70 - 30 degrees C. There may remain litre à small White flocculent precipitate of silver iodate, or, if as much as 3 - 3 mg. of HIO3 is present, this precipitate will settle out on cooling. 23. EXPERIMENTAL. Precipitation Methods. The first step here was to test qualitatively the completeness of precipitation with excess of periodate for the common cations. . The solutions used Were nitrates and approximately twentieth normal for precipitation reactions. The reagent used was a solution saturated With KIOA. at room temperature and diluted with an equal volume of water. This gave a rather dilute solution, but as most of the insoluble forms contain a high ratio of base, it was not usually necessary to add very large amounts to obtain an excess. Wherever --- necessary, however, solid K104 Was dissolved directly in the hot solution of the metal salt. The acidity or al- kalinity of the solutions was varied as noted in each case, with nitric acid or ammonium hydroxide. Ag. No precipitate obtained in presence of 2% by volume HNO3, sp. gr. 1.43. Even on adding metaperiodate solution to neutral Ag., NO3, considerable amounts of silver remain in solution, the amount increasing with the concentration of the silver nitrate. This effect is due to the precipitation reaction which liber- ates free nitric acid in forming the polybasic silver salt. No experiments were made with more alkaline solutions. Po" 2% HNO3--no precipitate. .25% Incomplete precipitation by test with K2CrO4. 24. -- - Mn -- 2% HNO3. Heavy black precipitate. ºr . Hg" —- 2% HNO3. ed. Coarse orange-red precipitate which settled immediately and filtered well. The filtrate was first reduced with SO2, the solution boiled to eliminate excess and treated with SnCl2. Only faint test was obtained. Hg' -- 2% HNO3. Behavior similar to Hg" except that pre- cipitate is lighter colored. Bi''' -- 3% HNO3. Precipitate filtered fairly well. Ohly slight test with H2S in filtrate in ammoniacal solution or in dilute HC1 after reducing with hydroxylamin. No precipitate Cu" – .35% HNO3. NH4OH solution + excese KIOA, then boiled till free of NH4, and neutral to litmus. Precipitate filtered 3 - very poorly, Filtrate showed no blue color with NH4OH. Cd" – .35% HNO3. No precipitate. NH4OH solution, same treatment as Cu". Precipitate filtered very poorly. Filtrate was free from Cd", by NH2OH.HCl, reduction and test With H2S. Fe''' – 2% HNO3. Precipitate filtered slowly. Slight test for Fe ' ' ' with ammonia and with KCMS, Cr' ' ' -- .25% HNO3. No precipitate oxidation occurrings Al''' -- .35% HNO3. No precipitate. 10% HNO3, Red precipitate and HMnO4 color. Neither pre- cipitation was at all complete. Zn -- .25% HNO3. No precipitates Ammoniacal solution + K104, then boiled until neutral to litmus. Precipitate filtered poorly. Filtrate free from Zn" - by reduction and tested with Has in neutral solution. Ni" -- NH4OH solution boiled as with zinc. Green gelatinous precipitate, filtered very poorly. Filtrate free from Ni, Co. -- NH4OH solution boiled as with zinc. Precipitate black and gelatinous. Filtrate contained cobalt. Apparently some oxidation of cobalt. º Ba' -- Solution just alkaline to litmus with NH4OH. Curdy precipitate, filtered well. Filtrate gives very slight test With H2SO4. Sr" -- Solution as for Ba”. Precipitate filtered fairly, well. Filtrate concentrated to 5 c.c. gave no test with Kaso.4° Ca." -- No precipitate in acid or neutral solution. Solu- tion slight excess NH4OH. Precipitate filtered poorly. Filtrate gave faint test with (NHºgao.4 after concentrating to 5 c. c. Mg" -- 1% by volume NH4OH, sp. gr. 90. Precipitate filter- ed poorly, Filtrate evaporated to 5 c. c. gave slight pre- --- - ----- * cipitate with (NH4)2 HP044 NH4OH). º --- *. *-* -- - - ----------- - - - º º --- - - - * - Iron. The precipitations which give most promise of usefulness are those occurring in acid solutions. Of these the reaction - from With ferrie iron, because of the marked distinction gº those occurring with aluminum and chromium, was selected for the first Worke A standard N_ solution of ferrie nitrate was made ea from electrolyºzer prepared in the laboratory. Preliminary experiments indicated a complete precipitation, on adding an excess of K194, With 3 c.c. Cr less of nitric acid, sp. gr. 1.42, per 100 c.c. of solution. * The filtrates were tested here by adding an excess of ammonia, and later experience showed that the last traces of iron are precipitated by ammonia only very slowly in the presence of periodates • An opalescence forms at first which is only coagulated completely on boiling for 1–2 hours. Further the character of the precipitate left much to be desired, and the method of adding the reagent as the slightly soluble Kioa, was not satisfactory. Accordingly periodic acid was prepared by the first method described in the discussion of reagents, Experiments to improve the physical properties of the precipitate showed that it was obtained in a form which fil- tered most readily, when the iron solution was itself added slowly to a hot concentrated solution of periodic acid. A new solution of ferric nitrate was made from Kahlbaum's analyzed salt, The solution was standardized by permanganate titration and by precipitation with ammonium hydroxide. The ammonia water used was a fresh preparation made by absorbing pure NH3, in water in ceresin bottles, which were sealed up immediately. All water used in this and subsequent gravimetric work was 27, tested for non-volatile residue. In the Work on iron the water was freshly redistilled using a block tin condenser. In some later work the laboratory supply being improved, this was found unnecessary. The standardization of the ferric nitrate solution, samples all drawn **tene se c.c. pipette and **** temperature, was as follows: Grams Fe2O3 1. ll 111 Average KMnO4 for 50 c, c titration 50c.c. sam- .3914 • 3915 • 3914 • 39143 ples. . Gravimetric 100 c.c. • 7828 • 7834 º • 3916 samples. - A large number of determinations were carried out in the effort to obtain a satisfactory precipitation of iron by this method. - A few will be given space here, to indicate the nature of the results obtained, - 20 c.c. of half molar Hsloe, 1 c.c. HNO3 per 100 c.c. of final volume of solution, and 50 c.c. of the above iron solu- tion were used in each case. Varying temperatures of precipi- tation, and varying temperatures and final volumes of the solu- tion at filtration were used, but without appreciable change. The following are typical results: - Found. - C# Fe...). Taken ſº pºecipitats In filtrate - 2 3 #mſ: lºs ºſiº l • 3915 . 3890 • OOll 11. • 3915 • 389.3 • OOO.9 28. The chief difficulty with the procedure is the ex- tremely fine state of division in which the precipitate is formed, When the solutions are mixed at room tempera- ture, no precipitate appears at once, but the nearly color- less ferric nitrate turns yellow and then deepers to a golden brown • After one or two minutes an opal escence Will appear. Thereafter the precipitate increases rapid- ly. It is probable that most of the iron found in the fil- trate passed the filter as suspended matter. The finest grained analytical papers were used for the above work. Nevertheless it was possible to observe a fine haze in the solution With strong illumination. Neither Munro nor Gooch crucibles of a practical density gave better results. Finally the marked adsorptive power of the precipitate Was shown by a determintation carried out with NaIO4. The mixture of this reagent in the presipitate made itself evi- dent during the ignition. The weight of the precipitate, .0165 mg high after first ignition slowly approached the norm- - al value on repeated heating in the full flame of the Meker ºne During this treatment a sublimate of sodium oxide was found on the cover up to the last ignitions - As a study in colloids this method was interesting. 29. --- - - - - - (, , ºn Precipitation reticas. º Potassium. º The other method of this type examined was the separation of potassium as the metaperiodate, KIO4. While the insolu- bility of this compound in Water is not of a quantitative order, the marked advantages that such a method would possess made it worth While to devise if possible, conditions which would give complete precipitation. The solubility of KTO4 at a series of temperatures from 0°––659 was first determined. Large crystals separated from Kahlbaum's preparation, previously described, were stirred With Walter for 4--.5 hourse This treatment was repeated until titration with arsenite showed a constant composition of the solution. The necessary temperatures were obtained with a thermostat controlled to within .05°C. For the determination at 0°C, the flask was immersed in a large ice bath. Tempera- tures were measured with a standardized thermometer. The titration of the dissolved material was made by the method of Müller and wegian, already described. - The arsenite solution was prepared standard by Weight from pure As293, according to procedure given in detail in the section on standard solutions for oxidation reactions. This solution was that used throughout the work on potassium. The buret used for the arsenite solution was of special design with a 25 c.c. bulb at top and 25 c.c. stem, the latter calibrated to twentieths of a cubic centimeter. The buret was clear back with graduation extending at least half way around the tube, which made it simple to bring the eye to the level of the meniscus. The opening of the glass tip was 30. contracted to conform with the Bureau of Standard directions for time of outflow. Both bulb and stem were calibrated by weighing the water delivered at a definite temperature. The corrections found were very slight, exceeding at no point • C3 c. c. , and in general from zero to .02 c.c. This burette was used for arsenite solution throughout the work on the determination of potassium. The solubilities of KIO4 in water, in grams per 100 c.c. of the solution are as follows: 0°C 259 359 450 .1546 .51.12 .785 i.ied 1.75 2.51 559 659 | ------ 2. - - rams per Solubility of KIO4 • The figure obtained at 25°C agrees fairly with that 10 a. 6 given by Rammelsburg,) but is lower than that given by sº 31. - The sp. gr. of the saturated solution at 25° is 1.00085. Means of reducing the solubility were to be found in the use of an excess of periodic acid, and of some proportion of an organic solvents - As the chief requirement of r any method for potassium Would be a separation from sodium, the conditions for the solu- bility of this periodate were studied simultaneously. The annyarous seize Kro 4 and Naro 4. crystallize accord- ing to Barker alike in the tetragonal system, though he does not mention having ºreºverals. - However, both Barker and Eºgree that Naro4.3H2O crystallizes in the rhombohedrial system and the latter states that the hydrated salt separated from solution below 30°C. The solubilities of Naro.4 in water and in periodic acid solutions, and of KIO4 in periodic acid solutions were neºtt founds The periodic acid for this purpose was obtained by the simple electrolytic method and recrystallized. Determi- nations of the nor-volatiºs made and corrections applied accordingly, As the excess of the free acid prevented the used of the volumetric method, the amounts of sodium and potag- sium were determined here gravimetrically. The samples were drawn with standard pipettes. These solutions, as also later the filtrates from potas- sium determinations, in which it was desired to determine sodium, or other consituent as the sulfate, were reduced as foºlows: - The solution being in a platinum or quartz dish and covered, a current of SO2 was passed in until all the iodine liberated at first had been reduced to H.I. The solution was then evaporated and ignited in a platinum crucible to Na2SO4, in the usual männers The SO2 was obtained from a tank of the liquid and washed thoroughly through a glass bead and water column, to remove any traces of non-volatile matter. The gravimetric determinations were certainly not as accurate as the titration method used above for the potassium salt in water, but sufficiently exact for these purposes. Standard tenth and hundredth molar periodic acid solu- tions were made by weight and checked by arsenite titration. The actual solutions were obtained by placing large crystals Of KIOA. and NaIO 4. polished with optician's rouge to a perfect fit. These were in Erlnmeyers with ground glass stoppers nearly filled with the periodic acid solutions, stoppered and sealed tight with resin and bees wax (2 to 1) mixture over the top. These flasks were piaeed on a rack, in a thermost at bath turning 50–60 R. P.M., which inverted the flask and then brought it - upright in each revolution. The time of this mixing varied from 4-5 hours for concentrated solutions to as long as 12 hours for some of the very dilute potassium solutions. Solubilities in Periodic Acid Solutions at 35°C. Grams per 100 c.c. of solution. __ –M- 100 Heroe 10 H5TO6 (Pure Water) KTo, • 476 - • 374. (.5112) NaIO4 19, 9 16, 3 (13.4) 2 º' some difficulty was encountered in obtaining satisfactory samples for NaIO4 solubilities in aqueous solution. The 33. anhydrous salt soon recrystallized as the hydrate and in a much more bulky condition. Hydrolysis of the salt probably did not take place in the presence of free acid. The results of the solubility determination nof Naſoa in pure Water indicate that this process occurred to a slight extent in neutral solution. Solubility of NaIO4 in water at 25°C. By weight of By titration Någsø4 of periodate. Gm - - Maſo.4 per 13.4 13s 8 100 3. c. soln. Next the solubilities in organic solvents were considered. Bºy, alsº most readily available material, and after that methyl alcºhol and acetone. Absolute ethyl alcohol was prepared from 95% by refluxing With an excess of freshly ignited calcium oxide for 3–4 hours and distillings This distillate was treated with metallic sodium equivalent to the remaining fraction of a per cent of water and distilled. This product indicated a content of 99.7- 99.8 by sp. ºr. measurements. When gravimetric determinations ed Were to be made this material was subject to simple distillation, rejecting about the last quarter, to purify at from traces of In Orl- volatile material. Methyl algºhol was dehydrated similarly, though of course, With less difficulty. - Acetone of the ordinary C. P. grade was distilled from calcium chloride, rejecting the first and last quarters. This product was treated with phosphorous pêntoxide and again dis-, 34. tilled, rejecting the last third, The question arises as to whether these organic sub- stances will be oxidized by periodate and give false results. For the simple metaperiodates it can be answered definitely that the reaction in no case proceeds in a degree to affect results during the time of these determinations. A solution of Naro4 in acetone, which is of these three the most suscept- ible to attack by periodate, was allowed to stand at 25°C for a week. At the end of this time no iodate could be detected by silver nitrate test as described in connection with the preparation of Hsſes. Similar tests were also made for each solution at the time of titration, but without indication of iodate. This, of course, may mean either that no oxidation occurred or that the iodates formed were too insoluble to give a test. In either oase, the solubility results are not affect- ed except in the possible event that the oxidation went far enough, with precipitation of the iodates, to actually affect the composition of the solvent. However, it is believed that oxidation if occurring at all in these solutions was slight. oxidation as found to occur in later work with solutions of the free acid could be readily detected by the odor of the aldehydes, but none was apparent here. These and all subsequent solubility determinations, with organic solvents present, were made in sealed flasks. The dry salts were first placed in the flasks and the necks constricted. Thereafter they were attached to the stem of a dropping funnel, and evacuated through the top of the funnel, to a few millimeters ed près Sures The stop cock of the funnel, which had been polish to 35. | a fit but not greased, was closed, and the funnel detached from the pump. The solvent was then poured into the funnel and admitted to the flask, without allowing air to pass the stop-cock. The constriction of the neck was sealed off at Oſlº 3 a. In this way explosion of the vapors in the gas space of the flask was avoided, This was especially necessary in dealing with the periodic acid alcohol solutions used later, The results obtained were as follows: Solubility of KIoa and NaIoa in organic solvents at 25°C, in grams per 100 c.c. of solutions Gºgoń CH3OH CH300 CH3 K 104 • OO 39 • O230 • Ol3 NaIoa , 349 l. 34. • 378 gm The Kre, in ethyl alcohol was titrated withi- Asso. \ - loo "8"|3 A check titration of the total iodine liberated in acid solutiºn calculated to .0030 gm KIOA. per 100 c.c. , The Naroa in ethyl º alcañol was determined gravimetrically. The methyl alcohol solutions were titrated. A gravimetric check on the NaIO4 gave 1.28 gms per 100 c.c. The acet one determinations were made gravimetrically. The ratios of solubility are for Kio, naro, Ethyl alcºhol ls 85, Methyl alcahol 1:54, Acetone l; 31. These facts alone remove methyl alcohol and acetone from sonsideration. In addi- tion to high insolubility of KIO4 in ethyl alcohol is strongly in its favor. - - Qualitative tests with glacial acetic acid show that NaIO4 as well as KIO4 is only slightly soluble in this material, 36. - The solubility KIO4 in methyl alcohol containing i: concentration of HgTog at 25° was found to be .0145 gm per 100 c.c. of solution. An attempt to find the effect of adding HETOs to ab– solute acet one solution was fruitless because of the prompt reduct; ion of the free acid by this solvent. This reaction Was studied in some detail as described in the second part of this paper. - Next the effect of dilution of the ethyl alcohol on the solubility of KIO4 was determined. - Solubility of KIOA at 259 per 100 c.c. of a mixture 90 parts C2H5OH + 10 parts H20 was found to be .030 gm. In the presence of free periodic acid there is undoubted- ly some oxidation of the alcohol. This reaction, however, is not as rapid as the statements of Langletºna Ramelº Would indicate. It is true that on standing several days at room temperature a hindredth molar solution of periodic acid in 85% alcohol becomes brown due to the liberation of iodine. Such a solution was alowed to stand ten days at about 20°C, then a portion evaporated rapidly to dryness on the steam bath, This residue would still oxidize manganous sulfate to permanganic acid in phosphoric acid solution, indicating that periodic acid had remained in the solution. The solutions from which the determinations's in periodic acid ethyl alcºhol mixture were made, were tested with silver nitrate, but without indication of the presence of iodate. 37. Solubilities in Alcºhol-Water-Periodic acid mixtures - ason in - at 35°C. These solutions are given in parts by volume. 85 pts Gºgoń 75 pts C2H5OH 15. ºpts H30 35 pts H2O *— Hero *— Hero Tº “5”8 To “5”6 gms gms KToa - - • OO90 • OO98 as K • OO15 • OOl? NaIO4 - 1.45 - 2. l8 as Na. • 156 • 232 Finally the solubilities in the 85 –– 15, Gaºgoń -- H20, –4– HEIOs at 0°C were determined, 100 NaIO4 .363 gm per 100 c.c. solution. KTO4 less than .5 mg " it m Evidently the cold solution is the most favorable don- ditiºn. With this data preliminary determinations were begun. Potassium was taken as the chloride from a fifth normal solution of the pure salt standard by Weight. - Periodic acid was also taken as a fifth molar solution. This solution carried about a 3 mg of non-volatile matter Weighed as sulfate per 10 c.c. The alcºhol used here carried 1 mg per 100 c.c. of non- volatile matter weighed as a sulfate. Preliminary determinátions: 1. 10 c.c.; Kol stirred with 15 °-9. Fºsſºs. 100 *-*. - 5 C2H5OH added with cooling to 0°C. Precipitate filtered, 38. washed free of Halo With C H5OH, IOA titrated. 33, 2 6 3 - N As2O3. Theory 40.00 c.c. Filtrate gave residue , 4 mg 10 - as sulfate, corrected for blank from reagents. II. 5 c. c. # KCl + 10 c. c. H5Tos- 85 c.c. C2H5OH M - 5 added, cooling to 0°. This gave final condition same as last solubility determination, 15.1 c.c. N. As203 Theory - 10 30.00 c.c. Filtrate gave residue .6 mg as sulfate, corrected. III. Same as II, except reagent added after dilution and Cooling. lá. 50 c.c. As2O3. Theory 30.00 c.c. Also titrated total iodine liberated in acid solution, required 58.00 c.c. N– LO Assºa • TV. Same as II., except .5 c. c. glacial acetic acid added be- fore diluting. 16.60 c.c. N. 1O As2O3. Theory 20.00 c.c. W. Same as TT. , except 1 c. c. nitric acid, sp. gr. 1-43 added, - - 17, 30 c. c. M As2O3. Theory 30.00 c.c. 1O - Filtrate gave. (.8 mg as sulfate, corrected. VI. 5 c.c. N. Kc1+ .5 c.c. HNO3 + 10 c.o. 1 molar N 5 Hero Diluted with 85 Cah OH. Precipitate removed as 6 * 5 before. 19, 80 c.c. N. As, O., . Theory 30.00 c.c. 15 3 3 Filtrate gave .3 mg residue as sulfates; corrected. WTT. 5 c. c. * KCl <+ , 3 c. C. HNO3 + 10 c.c. l molar H5Nos. 5 39. 19.8l c. c. – N As2O3. Theory 20100 c.c. - 10 Filtrate gave no residue as sulfate, corrected. VIII. Same as VII, except + .05 gm NaCl. 19.80 c.c. N. As 204. Theory 30.00 c.c. io “3-3 IX. Same as VIII except + 15 gm NaCl 19.85 c. c. N. As 20.4. Theory 20.00 c.c. The low results in I and II were evidently not due to failure of complete precipitation, as residue from concen- tration of the filtrates very little exceeded that from the reagents used. Neither was it due to reduction to the iolate 3.3 shown in III. Apparently this ascrepancy resulted from the precipitation of a basic salt. It was largest in III where the precipitation was made in the alcoholic solution. The addition of free acid in 17 and V decreased the ir- - regularity, but it was not until the periodic acid remaining in solution after precipitation had been increased to nearly tenth normal concentration for the final solution in VT that the results approached normal value. The final determinations were made with the following solutions and materials: Tenth normal sodium arsenite: --Six liters were made es- pecially for this work, standard by weight at 33. 0°C, accord- ing to directions given under section on oxidation reactions. Solution measured by burette previously described. Weight solution of Potassium Chloride, •0003 moles per gram--Kahlbaum's analyzed salt was dried five hours at 300°C. 40. A portion of this material was weighed out with 5 mg allowance for loss in fusion. This was fused in a small quartz lined electric resistance furnaces Weight after fusion, including vacuum correction, 14.9114 gms. This was dissolved and di- luted to 1000.00 gms. Samples taken with weight burets. . Sodium Chloride--Kahlbaum's analyzed——was fused in large platinum crucible in a quartz lined furnace and allowed to cool slowly. The outer layer of the crystals which had sep- arated first from the fused mass was used. Each sample was taken separately by direct weighings. Periodic acid--This preparation has already been des– cribed. It was recrystallized for certain determinations. Nitric Acid--'The C. P. grade was redistilled using quartz condenser and rejecting the first and last quarters. This product gave a non-volatile residue of . 4 mg per 100 c.c. Calcium Carbonate--Squibb's analyzed. Magnesium oxide—Kahlbaum's "zur analyse". Sulfuric acid--For the small amount of this used (15 c.c. per determination) the C. P. grade was considered sufficiently pure • Alcºhol--Absolute alcohol was prepared as described before. A blank of less than .1 mg per 100 c.c. of non-volatile matter weighed as sulfate was found. This was used as the last wash water to remove periodic acid. 95% Alcºhol--The labºratory stock of 95% alcºhol was refluxed for 3 hours with 10 gms per liter of NaOH, distilled, and redistilled alone, Blank of , l mg per 100 c.c. non-volatile tº e.e. was found. This was the alcohol used to complete 4l. precipitation. - A mechanical stirrer consisting of a single glass blade driven by a small motor was used to give a vigorous agitation of the solution especially during the time that the potassium precipitate was forming in the concentrated Water solution. The first six determinations were made by adding to the potassium solution in its original volume .35–.3 c.c. HNO3 and the sodiuſ; chloride, Where indicated, as a solid. 3. 38 gms of Hs 106 were dissolved in 5 c. c. water and this added to the solution, with stirring. After 2–3 minutes stirring, 10 c.c. of 95% alcºhol were added and the solution cooled by surrounding the beaker with ice. The remaining 80 c.c. of 95% alcohol to make a total volume of 100 c.c. were added in the course of a half hour. The solution was then allowed to stand lº–2 hours in ice. #1 –4 were filtered on paper. #5–6 were filtered on Munro crucibles. All were washed first 5–6 times with a solution 90 parts 95% alcºhol, 10 parts Water and iſ H5TOs, then with absolute alcohol intil 3 c. c. 1OO - of the washing gave no reaction with neutral Kl solution. This last was much more readily accomplished in the filtra- tion through crucibles. Both Walsh solutions Were cooled in ice to oºc. The precipitates were allowed to dry until free of alcohol either at room temperature or over a steam radiator. The bulk of the precipitate wº loose and Was transferred to the titration flask and dissolved in Water at about 50°C. The filters were washed with hot water, 80°- O - 90° and the washing added to the main solution. The whole 43. solution was then titrated in the usual manner for periodate. The algº hol filtrates and washings were evaporated to a small volume and reduced with S0s. The final evaporation of this strongly acid aqueous solution, d the expulsion of the excess HaSO were carried out in platinum. The sodium sulfate 4. residues were tested in each case for presence of free acid or of alkali from over ignition, with methyl orange and phenol- phthaleia. In each case the sulfate residues were neutral to both indicators. The periodic acid used in determinations 1-6 was the first crop of 150 gms from an electrolytic oxidation of 200 gms of iodic acid. 2, 38 gms reduced with soa, gave a non- volatile sulfate residue of 1.0 mg. (3.28 gm = IOC role H5106) Determinations of Potassium, No. KCl NaCl Kol Error Kol NaCl Error taken taken found KC1 In fil– found NaCl. gms gigs gms gms. trate gms gms - º - gms 1. .07485 .07430 -.00055 none 3. • O7474. • O7395 – 4 OOO79 • OOO1 2. .07393.0547 .07344 -.00049 • 6547 None 4. .07733 .0568 .07661 - .00072 • O569 + , OOOl 5. • O7425 • O'7381 - . OOO44 6. • O3402 • O3359 - , OOO43 The separation of potassium from sodium would appear to be accurate. Nevertheless the low results in the titration of potassium which were apparent in the last four preliminary experiments in a constant amount of ~15 c.c. to .30 c.c. of tenth normal arsenite solution or , 5 mg to . 75 mg of KCl con- 43. tinued through these determinations. No. 5 and 6 were made under exactly parallel conditions, filtering through Munro crucibles and indicated that this error is not a function of the amount of potassium taken. It may either be due to a slight but constant amount of KIO3 in the precipitate or to the precipitation of a small amount of the basic salt in the last stages of dilution with alcohol. In any case, this de- ficiency was so constalk that it seems justifiable to apply it as a corrections An effort to correct this error was made by increasing the amount of periodic acid used. At this point a new preparation of periodic acid was made by recrystalling 85 gms, from concentrated nitric acid. There still remained, however, a. non-volatile trace weighing , 6 - tº ºº is a - º ----------- - - ------ ºr ----- mg as sulfate per 3.38 gms of acid. -- In separate column here the results for potassium are corrected by a blank of . 15 c.c. = , 55 mg KCl. No. KCl H.- TO2, K61 Error KC1 Corrected KCl taken äseš found KCl Found error filtrate corrected - 7. .07776 3.42 .07717 –.0059 .07773 –.00004 .0001 8. .07453 5.56 . 07400 -.0053 .07755 +, OOCO2 . OOOl The increased periodic acid concentration evidently was Without effecte --- In the next determinations the addition of nitric acia Was omitted altogether. It was considered possible that low results were due to partial occlusion of potassium nitrates Increasing amounts of sodium were used in the series No. 9–13. Irregularities developed with the larger amounts. In No. 13 44. and 15 the original precipitates were redissolved and re- preprecipitated in the same manner. º sº. The following determinations were made as before with 3. 38 gms. of the recrystallized H5103 used in 7 and 8, in 100 c.c. volume. # KCl taken gm3 9 • O6835 iO . O'7540 ll...ozoag 13, . O7351 13 , 0.7305 14. oslav 15 s O744? 16 . O7936 NaCl. KCl taken Found . gms gms • 66%6 ... O ş68 . O'7485 .1543 .07896. • 15 . O'7799 • 2699 • O'?130 .3327 .08090 • 35 . O7374 , 3465 . O9065 Error KCl in NaCl. KCl Found KCl corrected gms • O6831 -, OOO14 , O7540 . OOOO , O7951 +. OOO3]. .07854 +.00003 .07175* +.00030 .08145 + .00018 . O7439* -. OOO18 • O9120 +, 0.1129 Error Filtrate Found NaCl. gms None gms gms • OS69 + , COOl • 1477 -. OO66 Not de- termin- ed • 2685 - . 0014 • 3345 – , 0.082 Not de- termin- ed • 33ll -. Olă4 * In these determinations, 13 and 15, which were redissolve- ed and repricipiated, a correction of .0003 gm KC1, obtained in a manner described later, should be added for the amount of KIO4 dissolved in the 100 c.c. of solution from which the sec- ond precipitation is made • # * * * * * * * * * * * * * * * * * * * * It was already evident when Nos. 11 and 13 were pre- cipitated that after the usual K104 crystals were completely thrown out, as far as could be detected by the eye, that on * standing the alcoholic solution precipitated after several hours, a second substance in coarse leaf-like form. 45. - But on titrating these precipitates, it was clear from the figures obtained that this material was not periodate and the rešults for potassium were unaffected. The titration of NaIO4 gave the same effect, with . 25 gm. of NaCl added. When 35 gm. of NaCl had been added in No. 16 the precipi- tate was finally contaminated with sodium periodate, and the result for potassium was high. - Meanwhile, in ll, 14, and even by a small amount in 13, Where the precipitate was reprecipitated, the results for sodium were low. It is not believed that the solubility of sodium chloride in 85% by volume alcohol, was exceeded here, but rather from the appearance of the crystals and their greater formation on standing that these consisted of NaIO3. This assumption is supported by the insolubility of this salt in alcohol, The addition of a little H2O2 to one of the filtrates from potassium gave immediately a white precipitate, also the solutions on simple standing at room temperature separated large crystals. The means taken to prevent this effect consisted in replacing the chlorine in the salts taken with nitrate, by evaporating with nitric acid. This step was carried out with all determin- ations described from here on. It had already been noticed that the solutions smelled faintly of chlorine soon after addi- tion of the H5106. Determinations 17–26 were carried out with the same volume relations and amount of H5106 as those just preceding: # KCl NaCl other Köl Error NaCl Error taken taken substanc- found KCl flound NaCl. gms gms es added correct- gms gms gms ed gms 17 - O7745 - 1575 • O'7761 +. OOO16 . 1574 – 2 OOOl 18 . O7904 . 1573 • O7903 -, OOOOl , lä73 it. OOOO 19 e O6886 s 3570 • O6896 f,000l.0 , 2569 –, OOOl 3O . O'731.1 s 2428 • O7317 E. OOOO6 - 2426 - « OOO2 21 - O'7731 , . 1217. 15cc • 07724 -. OOOO!? • O758 --. O459 H.; SO - 23 . O7814 - 1015 - 15 : * .07791 -.00029 .0859 -.oise HoS 33 s O776? .1% ºf .07772 -oooos MgO 34 . O7582 , 125 gº • O'7597 4s OOO15 Mg º 35 • O'538 • 18 gm • O7553 -, OOO15 36 s O7489 • 18 gm • O7474 +.00015 CaCO3 In Hos. 31 and 33, no HNO3 was added but the mixed chlo- rides were evaporated with In the HaSO4 to fumes of S03. the precipitation, the separation of feathery translucent crystals, probably NaHS04, occurred on allowing the alcoholic solution to stand at 6°C. A titration of the total iodine liberated in acid solution by the precipitate in 26 indicated that about .5 mg of Cao was Carried down as the iodate. A similiar titration of 34 show- ed no Mg (103) present. It only remained to determine, if possible, the nature of the constant correction, which it had been found desirable to apply to the results obtained for potassium by the periodate titrations For this purpose two large samples were taken and the precipitates, after washing 47. than one-half, as actually observed, of the total error is to with absolute alcohol, dried in vacuum over H2SO4 to constant Weight, and then at 105°-110°C, Where no change occurred, The results were as follows: - # KC1 KIO4 gms KCl gms by titration Error taken Theoret- Found gm gms ical Uncorrected Corrected KCl 27 .21795 - 37235 .67235 .31734 - 31789 –.00006 - | 28 - 21413 - 66040 - 66045 - 31336 • 21391 -- 00021 Also two small samples were determined in which both periodate iodine and total iodine liberated in acid solution Were titrated: KC1. Gms KCl by titration of Error KCl # taken Periodate total by per— total gms Uncorrected corrected iodineſ iodate, iodine 39 . O1429 • O1370 • Olá35 • 61398 ~00059 – .00031 30 e Old 44 • Cºº?? • Olo:32 • Olol.4 • OOO67 -, OOO3O The error by total iodine titration is in each case close to one half of that by periodate titration. The presump- tion is that the difference arises from-the presence of KIO3 in the precipitate. Had this potassium been present as K104, the results of the total iodine titration would have been greater by one-third of this difference, or two-thirds, rather be ascribed to this deficiency in oxygen. . The remaining dif- ference amounting to .18 mg. of KCl or ,05, c. c. N. arsenious - - 10 acid, in the periodate titration on the average determination 48. is probably due to solubility. The above titration of periodate and total iodine was carefully checked with the same solutions and manipulation CIl 3, sample of Kahlbaum's K104. Here the total iodine was exactly four times the periodate iodine, to . Ol c. c. in 76.00 c.c. Finally an effort was made to evaluate the solubility effect. In earlier work, though estimates had been made of the potassium remaining in the filtrate, the non-volatile blank on the reagents amounting to about 1 mg. With slight contami- nation of iron from the platimum dishes, had left these determin- &tions in question by , 2-, 3 mg. - In the work just described, the reagents, in particular the periodic acid, had been purified again, until the blank on these amounted to , 3 to , 4 mg. The filtrates from the determinations were evaporated in quartz, and gave residues which differed from the non-volatile blank by # scarcely weighable emounts, .05-. 1 mg. sulfate. It was decided to precipitate periodate from solutions of the same alcohol water ratios and concentration of H5106, but of 300 c.c., total volume. # KC1 KCl found, gms by titration ºf Error by taken periodate Total Periodate gms Uncorrected Corrected Iodine I di Uncorv Cor- To- - - rected rected tal 31 - O'7845 - O'7748 • Q7803 • 7781 • COO97 . OOO43 ... O C34 32 - O'74Ol. ... O73OO ... • O'7355 ... • OO1001 e OOO56 Plainly the error is not proportional to the total volume, increasing only 75%. If this increase represents only 49. detected. However, if iodate had taken the place of periodate the amount of the precipitate held in solution in the 200 c.c. by which the total volume was larger here than in previous determinations, then one-half of this difference in error, or **-ā-º * .38 mg. is the solubility correction to be ap- plied per 100 c.c. of solutions The remaining error is made up of the amount dissolved in about 20 c.c. of washings, part- ly absolute alcohol at 9°C, which is believed to be negligible, plus the constant correction as determined in No. 29 and 30, as due to contamination of iodate. º. The results obtained in 27 and 28, wherein the weight of the precipitate was theoretical, and those in 31 and 32, Wherein an increase of the volume threefold, without any distinct. increase, in the amount of the constant error, and preclude the possibility that this last might be due to the separation of a small quantity of basic salt from the final alcoholic solution. Taking the average amount of this eeror as the equivalent of • 1 c. c. of N arsenious acid or tºo-thirds of the total cor- rection of #s c.c. for the periodate titration, it would corres- pond to a deficiency of I.20% of .9 mg. There was no such vari- ation from the theoretical weight for pure KIO4. Even replace- ment of periodate by nitrate in the precipitate would have pro- duc ed a lowering in weight by .65 mg., which also would have been º the difference in weight to correspond to .1 c.o. of N. As2O3 10 would only be .08 mg. Hence the total correction of . 15 c.c. of N arsenious 10 acid, equivalent to .55 mg. of KCl, for precipitation made from 100 c.c. total volume, is analyzed into .05 c.c. due to, solu- bility of KIO4, and • 10 c.c. due to the presence of iodate in the precipitate. The procedure recommended for the determination of potassium in the presence of sodium, calcium, magnesium, or S03 is as follows: Evaporate the sample containing not over • 3 gms of KC1, with HNO3 to dryness. If an excess of H2SO4 is present volatilize this until only the bisulfates remain, or better ig- nite the neutral sulfates. Dissolve the residue in 5 c.o. water, stir vigorously, and add 3.38 gm. Hs 106 dissolved in 5 c.c. of water. Continue to stir until no further crystallization of KIO4 CC Cliº Se The solution should be between 159 and 20° at this time. Add 10 c.c. 95% alcohol, with stirring. Cool to 0°, and add in the course of 3 hr., 80 c.c. more 95% alco- hol to the cold solution, stirring 3–3 minutes with each ad- dition of 10–30 c.c. Let solution stand at 0° for 14 hours. Filter on a Munro or Gooch crucible. Wash 5–6 times with 85 cºisoH -15H30 Tizi- H5IO6 (.35 gm per 100 c.c.) at o°C, and finally with absolute alcohol at 0° until no appreciable iodine color appears on adding a few drops of 5% Kl solution to 2-3 c.c. of the Washing. A faint color may appear on acidifying with HCl, but this should be entirely removed with 1 drop N. Na2S2O3. Dry the precipitate at 70-75°, tap the main º into a 350 c.c. Erlenmeyer and dissolved in 100- 150 c.c. water at 500-70°C. Wash the crucibles and precipi- tation beakers with water at 80°-90°, and add to the main solution. 51. Add at least l gram of borax and 1 gram of boric acid, for each 10 c.c. of N arsenaeusºs in the titration, swirl to solution of ºss, a cool to 20°C. Add at least 1 c.c. of 5% Kl solution for each cubic centimeter of ar- senious acid to be used, titrate with N arsenite, adding to 10 the volume used . 15 c.c. for solubility and iodate correction. c.c. N arsenious acid (corrected) x .003738 = gms KGl. 1O - - Oxidation Methods. The limited observations of earlier investigators on the oxidation reactions of periodate have been mentioned fully. Mäny of these have concerned reduction by organic compounds, but determinations of the latter have not been considered to lie Within the scope of this work. One such reaction, which was encountered in the course of other work will be mentioned. Th & inorganic compounds, whose oxidation has been described in the literature, are ferrous, manganous and cobaltous salts, iodides, iodine, hydrochloric and hydrobromic acids. In addition, chromic salts have been found to be oxidized to chromates in neutral or slightly acid solutions, vanadyl salts to vanadic acid in acid or neutral solution, and cerous salts to ceric salts in neytral or slightly acid solutions. The question at once arises whether these reactions are quantitative. - But further, and presenting greater difficulty is the question, by what means is the completion of the reaction, if quantitative, to be indicated: These will be of two kinds, first the use of some sharp effect, i. e- a direct end point produced by a negligible ex- cess of the standard periodate solution, second, the use of a secondary titration to determine a finite excess of the periodate solution, Whether some means is found to define the completion of a reaction will determine largely the possibility of its becoming the basis of a new method. Hence, the work described in this section will be arranged according as the method under consideration utilizes a direct or a secondary end-point, 53. Standard Solutions. For methods using direct end-points, only the single standard solution of the periodate will be required. For meth- ods of the other type, there will be needed, in addition, 3. standard solution of some reducing agent. For all work car- ried out here the latter will be arsenious acid. Both solu- tions Will be taken up here, before any methods are discussed, particularly since the arsenious acid solution was made the fundamental standard for all volumetric works To establish this standard a start was made with a preparation of Kahlbaum's analyzed arsenious oxide, which was certified to be 99.98% pure. This material was subjected to careful sublimation in a slow stream of pure dry cargon dioxide. A number of separately weighed samples of the resultant product were dissolved and titrated against a standard iodine solution. Within the same hour, and under uniform conditions, samples of the original preparation were dissolved and titrated. Although samples requiring from 60 c.c. to 130 c.c. of the iodine solution were taken no difference between the two materials was revealed. Hence the original material was considered a sufficiently accur- at e volumetric standard. The arsenious acid solution was made by placing in a Weighed weighing bottle a definite amount of the Kahlbaum's oxide, dried 2–3 hours at llo” and reweighed. The drying loss was always less than .02% and could be anticipated, so that, correcting also for .02% impurity, and to weight in vacuo, it was possible to choose the first amount taken to give exactly the calculated weight of the dried material for the solution 54. required, However, in no case was the step of drying and reweighing omitted. A known weight of the dried material, in general slightly over 39.687 gms was obtained for each prep- aration of tenth-normal arsenious acid, to be diluted finally to volume in a calibrated flask of 5.99963 liters content. * - Five such solutions were prepared and used in all in-ºre +- vestigations embodied in tº thesis. In each the calculated correction factor differed from unity by a negligible amount. The arsenious oxide was brought into solution in the fol— lowing manner. The sample was transferred to a 3 liter Jena flask, 60 gms of Kahlbaum's "zur analyse" sodium bicarbonate were added, about 150 c.c. of water and the mixture swirled until the oxide was entirely moistened Then water was added gradually and the solution heated on the steam bath to com— plete solution in 700–800 c, c. total volume. All water used in the preparation had been freshly boiled and cooled while passing through it a vigorous current of carbon dioxide. Fur- ther the entire process of solution of the oxide was carried out in an atmosphere of carbon dioxide. This concentrated solution of the oxide was transferred to the standard flask, diluted to volume at a known temperature within a few degrees of 20°C, and then transferred to a large glass bottle, all with– out exposing it to the air. In the final container the gas space above the solution was filled With carbon dioxide, and all solution used, by an automatic arrangement was replaced by the same gass Solutions so made and protected were found to be entirely 55, stable for the purposes of this work. One such was examined two years after first made and the variation from the original concentration was less than .1%. In other cases it was not possible to detect a change in the reducing power during the time any one solution was in use.”ºre in general a period of three to six months. - Based on this standard, there were prepared a number of standard solutions of periodates. These were of these types, metaperiodate solutions, dibasic alkali periodate solutions, and periodic acid solutions. Metaperiodate solutions were made only for preliminary tests of stability and were soon found unsatisfactory. A solution of KIO4 or NaIO4 in Water at room temperature in a very few minutes gives a distinct odor of ozone. If a concentrated solution is allowed to stand in a stoppered bot- tle for several days this odor becomes very strong. This spon- taneous decomposition appears to be steady. Interesting to note in connection with the discussion between Garzarolli-Thurn- lsº &nd W. ause: who both tried to prepare periodate by the action of ozone on iodides or iodates, and as the latter con- cluded in 1912, with negative results, is that this decomposition, the reverse reaction, apparently does not proceed to an equili- brium. The accompanying diagram is taken from the results of titration of a tenth-normal, i.e., twentieth molar solution of sodium metaperiodate, without any other addition, starting from the time of preparation. This work was done in July and º the average temperature in the laboratory was 389–30°C. The 56. decomposition is seen to be practically a linear function of the times - H H- + - TTTTT Decomposition of metaperiodate tenth normalesolution at 30°. Amºs, Mica. Fortunately, it was found that this decomposition could be very greatly retarded, if not completely prevented, by the addition of one equivalent of either free alkali or acid. In the first case, there is obtained the dibasic di- stºricasts, K2H310s. Only the potassium salt can be uti- lized here, since the corresponding sodium compound is very in soluble. These standard Kºhg106 solutions were used largely for one type of method, and were prepared in the following manner, A definite weight of dried KIO4 was obtained in essentially the same fashion as described for As293. This was dissolved by adding an equivalent amount and 10% excess of KOH, taken as Merck's C. P., of a determined free alkali content, and heating in a 3 liter Jena flask on the steam bath. This solution was &c complished quickly to prevent attack on the glass vessel and consequent precipitation of basic periodates. Such a solution could be made practically standard by weight. The accompanying diagram indicates the rate of decompo- sition of a tenth-normal K2H3106 solution. It is certain that not even the small decrease in oxidizing power shown is due entirely to decomposition of the periodate, for in every case there was a precipitation of traces of periodates of iron and other bases, which unavoidably came into the solution from the KOH and from the glass bottle. This material settled very slowly. Such a solution gives a gery slight odor of ozone. TTTTT *-i- Hºrrºr Tºº º H H-T- -I-T- Decomposition of tenth normal K2H310s, Fabout—35°Gº Ass Asso, Micº. 58. The third type of standard periodate solution used is one of periodic acid, or what serves the purpose quite as well, sodium metaperiodate acidified with one equivalent of sulfuric acid. Such a solution is stable both in regard to decomposi- tion and to precipitation reactions, While the decomposition of such solutions was not followed as carefully as that of Kºsſos solutions, they are certainly satisfactory in this re- gard. One such solution gave perfectly constant titrations with arsenite through a period of four months, and then kept in the dark and titrated two years later showed no loss in oxidizing power. Such a solution does, however, give an odor of ozone, a fact apparently paradoxical to its constancy of titer. Unfortunately no satisfactory methods were devised based - on oxidations in acid solution, for which such a type of stand- ard reagent would be most valuable. Before proceeding to specific methods another oxidation reaction, that of periodic acid with acetone, will be described briefly. On adding crystals of H5108 to pure anhydrous acetone solution, the solid starts to dissolve, but within l–3 minutes at room temperature, about five minutes at Ö9C, a cloudiness appears throughout the liquid. This increases rapidly to a precipitate, finally separating in a white curdy condition. If there is only a small excess of acetone present, the reaction may become quite violent. It was found that the precipitate consisted of practical- ly pure HIO3. Other products whose presence was definitely established were acetic acid and considerable amounts of formaldehye. 59. No evidence of oxidation beyond formaldehye could be found, either by test for formic acid, or by calculation from the titration, in a quantitative experiment of the total acid liberated. Also no evolution of carbon dioxide was observed. Chromium and Vanadium. The determination and separation of these two elements by differential oxidation with periodic acid appeared at the first a very promising possibility. But it was quickly dis- covered that it would be necessary to devise some direct end- point in acid solution before the volumetric separation could be accomplished. - The oxidation of either proceeds readily in neutral so- lution. In even moderately acid solution, the chromium is oxidized very slowly, if at all, while the vanadium is still oxidized readily. In strong acid solution the vanadium it- self is probably not oxidized completely. The direct bearing of these conditions on the matter of the end-point is that, although vanadium may be oxidized in the presence of sulfuric acid, While chromium remains unattacked, an excess of periodate can not be determined here in the usual way after neutralizing with borax. For such an excess would immediately react with the chromium in the borax solution. - These facts were shown by the following preliminary experiments. First standard solutions of chromic and vanadyl sulfates were prepared. . 60. Starting with K2Cr207, recrystallized from the C. P. grade in conductivity water, an exºct amount of this salt Was taken to make 6 liters of tenth normal dichromate for oxidizing purposes in the general manner previously described for Assºg' This salt was dissolved in half a liter of water, acidified slightly with H2SO4, reduced completely with S02, and the excese of SO2 removed by passing a current of CO2 through the boiling solution until no test for S0s was obfained with starch-iodic acid paper over the mouth of the flask. In precisely a similar fashion a twentieth normal solu- tion of vana.dyl sulfate was prepared, starting with NH4 V03. This solution was kept saturated with CO2 during the prepara- tion. The latter solution was also titrated against , Kºnoa, which in turn had been standardized against sodium cºst. obtained from the Bureau of Standards. This standardization also showed the vana.dyl sulfate solution to be exactly, - With these solutions the following preliminary titrations were carried out. I. To 50 c.c. of N. chromium solution were added 59.92 c.c. 10 N. K2H3106 solutions This mixture was heated to a boil, kept 1C at 80–90° for five minutes, cooled, treated with borax, boric acid and KI, and titrated as usual with # As293. Required 9.95 c. c. Theoretical 9, 93 c. c. The titration was duplicated with the same result. II. To 50 c.c. n venadyl sulfate were added 30 c.c. of N Kzºg, - 30 - 1O ". 5 c. c. 5 N/H2804, and 100 c.c. H20: The sºlution was heated to 40°, allowed to stand 3–3 minutes, neutralized and titrated for periodate. Apparently some compound of periodate is formed here which reacts a little slowly with KI. But, if the iodide is allowed to react for 2–3 minutes, the end-point is definite and normal. Required 5.02 c.c. N. As2O3. Theoretical 5.00 c.c. Finally an *. titrate the vanadium in presence of chromium gave no result. It now became essential to find a direct end-point for periodate. Two general possibilities were investigated but neither with success. First a number of dyes were tested for a sharp decolor- ization in dilute solution on addition of a few drops of tenth normal periodáte solution. - Second, an attempt was made to distinguish an end-point by the rise in oxidation potential against a saicmel. electrode produced by a solution of periodate. The results obtained in this work were anomalous and are as yet unexplained. It was carried on altogether in an at- tempt to obtain a satisfactory end-point in the oxidation of pure vana.dyl sulfate with periodic acid. The vanadyl sulfate solution was the one just described. The periodate solution was one of the third class, made by dissolving an exact weight of pure NaIO4 and adding one equivalent of sulfuric acid. The NaIO4 had been analyzed for T207 content. The solution was taken as standard by weights The best conditions that were found were as follows: Acid concentration, about 5 c. c. H2SO4 sp. gr. l. 84 per 100 c.c. of solution. A large concentration makes the ‘gogºv NNV '83"I'IHSX1003-ºl HVAA HORIO'HO • Uſoțqntos eq 'epoț¢rºdi pareçueſ, e qį Įſ, e qieg Ing TÁpētī£ A JO Uſoļº， īſą ſą oȚI? etuo Iſ; oºTTI 63. rise in potential at the end-point less abrupt; a smaller Concentration begins to permit oxidation of the vanadyl salt by iodic acid, thus rendering the reaction inexact. Temperature. Change in temperature as the diagram in- dicates, had very slight effect on the form of the curve or the voltage values obtained. The real difficulty with the method lay in the extreme slowness with which the potential assumes a constant value at the end-point. Though the voltages as finally obtained and plotted show an excellent break, it required in general from 80 mintues to 3 hours to obtain each of the points after the theoretical end-point had been approached, Increasing the temperature of the solution improved this difficulty slightly. The addition of 1 gm MnSO4 to the solu- tion was rather more effective, But neither brought the time required within practical limits. The Halogens. There remain to be describeč a series of determinations and separations of the halogens. From the funadamental nature of the reactions utilized, it is necessary in each case to use an excess of periodate. This is always determined by titration with arsenious acid in borax-boric acid solution, Hence for this work a direct end-point would be of no value, even if it were possible. Iodines The work of previous investigators's on the reaction KI + 3KI04-24KI03 63. has already been mentioned, There are two main difficulties which lie in the way of its adaptationto a determination of iodine. First, since the initial product of the reaction is free iodine, and since the solution must be hot, there will be danger of loss of iodine by volatilization. Second, since the reaction in general proceeds to completion rather slowly, there will be danger of decomposition ºf the residual periodate. The first difficulty will be remedied by allowing the oxidazoºke place in a closed flask, and by having the solution; somewhat alkaline • Müller has shown that an excess of alkali retards the reaction. On the other hand, as was first shown by preliminary experiments, the reaction also goes more slowly in acid solu- tion than in neutral. Extended observations on the reaction in acid solution will be given later. It will suffice to say here that the most satisfactory conditions for carrying out the reaction for a determination, were to be obtained by mixing a solution of the neutral alkali iodide With an excess of N K2H3106 solution. No other reagent was found necess&ry - 10 for this steps Two standard solutions of potassium iodide, each #enth - normal for precipitation reactions, were prepared for this Work. The first was used for the preliminary determinations, the experiments on the rate of reaction of KI and K2H3106, and for some of the final determinations on pure iodide. After some time a very faint coloration of iodine was noticed in this solution and the second solution was made and substi- 64. tuted at once, Both solutions were kept under an atmosphere of hydrogen, and the liberation of iodine in the first was traced to the presence of minute quantities of nitrogen oxides. These came from the silver nitrate solution used to purify the hydrogen. With the second solution silver perchlorate re- placed the nitrate as a wash solution for the hydrogen. The solutions were prepared in precisely the same manner. Potassium iodide was recrystallized in carefully purified water from the chemically pure salt. The product was found essen- tailly free of chloride, bromide or iodate. No effort Was made to prepare these solutions of exact tenth normal content by weight of K.I taken. Further they were not used as volu- metric, but as weight solutions. Accordingly they were stan- dardized by taking samples of known weight and precipitating with silver nitrate. The following values were obtained, in grams of iodine, weight corrected to vacuum, per gram in air of the solution; A B C Average Solution I • O126555 , C.126569 e Cl26533 • C126553 Solution; II • Ol25961 s C135965 • Cl25966 . Cl35964 The most important consideration that remained to be examined was the time required to complete the reaction, and in particular the effect of temperature on this. The simplest qualitative experiments had revealed that at room temperature with equivalent amounts the time would be infinite. As it happened the reaction could be followed accurately by mixing the reagents in a known volume, at succesive periods pipetting 65. out aliquot portions and titrating the residue periodate. In this manner a series of experiments were made at 250, 35°, 45° and 55°C. In each case a weight of potassium iodide solution to correspond to l mole, and 150 c.c. of N. K2H3106, or 3 moles, ***. These quantities re-ºvalent re-ºries. conversion to iodate. The perio- date solution was placed in a 350 c.c. standard flask, and immers- ëd in a thermostat bath of the required temperature for the particular experiment. Similarly the iodide solution was brought to temperature in a separate flasks Then the iodide was ad- ded to the periodate and diluted immediately to 250 c.c. As was always the case on mixing these solutions, a small amount of free iodine was liberated at the start giving a deep orange colorations As the temperature was increased with successive experiment the color appeared more sharply at first and dis– appeared sooner. The point in each experiment at which the color was found to have practically disappeared, is indicated by a large circle in the diagram. This is not a sharp change, however, at any temperature • At the times required to give well defined curves, 35 c. c. samples were drawn, neutralized with borax, treated with an excess of KI solution, and titrated with N arsenious acid. this titration gives periodate existing : such, plus the e- quivalent oxidizing power of such periodate as has been used only to oxidize iodide to free iodine. A second experiment was made at 359C, with the same abso- lute amounts, but diluted to 500 c.c., and followed with 50 c. c. *ITISSIOog-MHVAA Hººſoº) - Fºogºv NNW oºH 2 + HOXI 2 + 2OIX® IXI +ºoIº Hºx2 -: uoſ įoeeg 66. --------- samples. Thus the curve shown is comparable with that of the first experiment. These results led to no theoretical formulation. The only clue obtained her 6 to the order of the reaction is the fact that dilution, in the experiment at 25°, gave no marked change in the rate. This fact would denote a reaction of the first order. However, the empirical curves indicate plainly that the high temperatures and an excess of periodate will be required to obtain quickly a complete oxidation of the iodide to iodate. All determinations of iodine is were carried out by the same general procedure. 50cc. of N/10 K2H310s were introduced to 3. glass stoppered 250cc Erlenmeyer flask. A flat ended glass capsule, containing the iodide solution, 3–8 cc in volume, was lowered into the flask, placed upright on the bottom. Thus the two solutions were kept separate, until the whole system was heated to the temperature at which it was desired to run the experiment. - lil 3, i When this was reached the flask was stoppered, then tilte m&nner to upset the C&psule and the solutions mixed. The mixture he period and at the temperature noted in the Was heated, for tables, then cooled, and titrated. In certain experiments where the determination in presence of a very large amount of bromide was to be tested, only 35 cc of M/10Kºiz Ios were used, and 75 cc H20 were added. The bromide was #g always introduced by dissolving solid KBr in the iodide solution. The KBr used was a preparation of especial purity made in this laboratory for use in atomic weight aeterminations. This is the same material that was used later for standard N/5 KBr, before 67. No Water Was added in the reaction mixture in determinations 1 – 22. In later experiments the effect of dilution was determined and larger flasks were required. First a set of determinations were run with varying excess of periodate in order t to find the time required to complete the the temperatur -- in d 3. reaction. These Were heated on a steam bath an the flask was about 80°C. 1. # Iodine Excess Time heat- Iodine found. taken grºs K2H3102 ed. min. gºls. - l .07273 14. 3 io .0725 – .. 3 a .97e, 12. 41 7 . O'737 O. O. 3 . O'798. 1C. 90 5 . O'945 — . 4. 4 . O’89 ll. 3 C, 5 . O'791 + .. 3 5 .0838 17. 5C 3 ... O & 37 —l. 1 3 . O375 7. 32 3 .0832 –4. 3 7 . O966 2.95 not taken .0965 — - 1 3 - - - - iodine volatilized 9 . O964 3. O9 35 • O2&l — .. 3 10 : 1008 , 9'? 5C . 1011 + .. 3 11 : 1.OC'? 1. Cº. 40 • 1005 – , 3 12 - 1004 1, 16 3O ... 1003 - . 1 The above experiments were made at the same time as the rāte of reaction experiments ana before the importance of temp— erature control in obtaining complete reaction was realized. In the following determinations the temperature was watched care- fully. In 13 and 14 a temperature of 709 was maintained. In the other experiments of this set and in all determinations of iodine hereafter, the flask was immersed in a steam bath and the temper- 68. ature of the solution was 96 – 98° Centigrade. Also in these determinations, after the first color of the iodine had disappeared, the solution was neutralized with boric acid, and then heated until the iodine liberated by the acid had also disappeared. # Iodine Excess Time heat— Iodine found Error mg of taken gms Kº H., IO2 ed. min. glū8 . iodine - cá.3 º - 13 . C964 3.12 35 – 70°6 .0962 — .. 3 14. . Cºº?? 2.48 30 – 70°ge .0977 0.0 15 - 1009 . 95 40 ... 1008 - - . 1 16 . 1008 .96 40 , 1009 + -l l? ... — - - – - determination . - lost, flask cracked 18 . 1003 . 31. 70 ..1002 O. O. 19 .0998; 1.51 22 . O998O 0.0 3C . O996 1.84 21 .0996 O. C. In determinations 21 – 36 KBr was added as noted. In this C 3, 3 & there is 8. possibility that the reaction KIOA + 3KBrº-H2O > KIO3 + 3Br + 2KOH may occur to a slight extent. The bromine liberated Will, however, when treated with excess of KI in the final titration, liberate a quantity of iodine exactly equivalent to the periodate used in the above I. eaction. So the reaction can theoretically, proceed without affecting the results obtained for iodine. It will be necessary, however, to see that none of the bromine liberated is volatilized. Accordingly, it will not be advisable to open the flask during the reaction to &ºid boric acid, and in fact, this step was oſſmitted in the following determinations. In 31 and 33 the usual pro- ceedure was used. In 33 - 36, 85 cc M/10Kºhg10s + 75 co H2O were used. 69, # Iodine KBr gºs Excess Time heat— Iodine Error mg taken K2H310s C.C. ed -- found gºs iodine. 31 . O982 ... 6 1. 38 15 .0983 O. O. 32 . O965 , & 3. 13 1C) . O964 –0. l. 23 . O38C 5. O 6. 35 3 . O38C, - O. C. 34 . O383 5. O 6. 15 3 ... O 383 +0. l 25 . O386 5. C. 5.95 3 ... O 383 O. O. 26 , C415 5. C. 4.69 s .0414 –0. l. The presence of the bromide facilitates, if anything, the oxid- "- ºr - ation of the iodine and cons ë É it, quently the Whole determination . The estimation of small quantities of iodine in presence of + considerable amounts of bromine appears to be particularly good. The following experiments 37 – 30 were made with pure iodide solution to determine the effect of dilution:- # Iodine Excess Total vol. Time heat- Ioane Error mg taken K2H310s ºf ºn ed, min. found of iodine. 37 . C988 3. 32 100 15 - C390 + , 2 38 . O985 3. 38 15C 15 . O93? 4. .. 3 29 . O985 3. 38 200 15 .0985 0.0 30 . O98? 3. 28 250 is . O987 0.0 Series mumber 31 – 33 and 34 – 36 were rºadie at 350 cc dilution, with decreasing amounts of excess periodate to determine the lower limits in this regard, at which oxidation would be complete in definite times of heating. # Iodine Excess Total vol. Time heat— Iodine Error mg taken K3H310s of solution ed., min. found of iodine. G G . - 31 .0785 13. 78 35C 8 . O'? 34 - . 1 3 2 . O350 G. ºl - 250 8 . O350 O. C. 7Os 33 . O923 3. 33 350 8 . C91.3 —l l 34 . O'704 13. 35 350 5 min . O’C4 C. C. 35 - O'76O 14. ÖC) 350 5 min . O'754 – .. 6 36 .083. 10.63 350 5 min . O814 —l. 7 In conclusion, the chief error to be avoided is incomplete ving low results. i i. oxidatiºn of the iodide iodine, g Decomposition of the periodate is slight at the temperature of the steam bath, and will not affect, determinations which do jot require over 15 – 30 minutes for oxidation. More accurate results are obtained by use of a large excess of periodate than by permitting the reaction to run over 15 minutes. Numbers l, 31 and 33 made with 10 – 14 cc excess N/10K2H310s and heated 8 – 10 minutes show no tendency to high results from decomposition of the periodate. Completion of the reaction is not retarded by dilution, and is accelerated, if affected at: all, by the presence of bromide. The procedure recommended for determination of iodide alone or in the presence of promide, is as follows :- Take 10 to 25 percent excess N/10 K2H3106 above the amount required to oxidize R.' I to R' IOE in the sample containing iodine. Dissolve the iodide sample in 5 – 10 cc water. Placing the two solutions in the flask and capsule as described, heat to 979 – 98°, stopper flask and mix solutions. Heat now for lò – 12 minutes. No color of iodine should remain after 5 minutes heating. The persistence of iodine color indic sites too sººn excess of periodate. When heating is completed, remove stopper from flask, cool, neutralize with an excess of boric acid. No color of iodine should appear at this --- ºil & point, this being an accurate indication of complete oxidation. However, with bromide content higher than 75% of the sample, a faint coloration of bromine may appear slowly after adding boric acid. Any color of iodine will appear at once. After the addition of boric acid, add 5% KI solution, at least 1 c.c. for each cubic centimeter of N/10 As2O3 used in the subsequent titration, c.c. N/10 K2H3106 reduced x .0021153 = gm. Iodine. Bromines The liberation of free bromine from solutions of bromide ion by periodate, occurs very slightly in neutral solution. In acid solutions the reaction may be carried to completion if some means of removing the bromine from solution be found. If this can be done without causing decomposition of the excess of periodate in solution, the latter may be titrated and the reaction can be used for a determination of brömine, - It was not considered advisable to boil solutions of periodic acid, to expe. bromine , , over the gas flame. Instead, a flask was arranged whereby a vigorous stream of air could be drawn through the sºlution, while immersed in the steam bath. This sº was fitted with a ground glass stopper taken from e g3.3 washing bottle and provided with inlet and exit tubes. The tube leading down into the solution was sealed off, blown out to a “small bulb, and the latter perforated with a number of small holes about 4 mm. in diameter. When the other tube leading out from the top of the wash bottle was connected to an aspirator, a stream of air was drawn through the solution and the bromine removed as it was formed. The solutions and reagents prepared especially for this work were as follows: 72, N/5 potassium bromide solution was prepared accurately by Weight. The salt was that already used in the separation of iodine and bromine. It had been prepared by the reaction of recrystallized K2C204 and especially purified bromine, removed from solution and recrystallized, then fused in platinum. A test for chlorine with a nephelometer, showed less than .03 mg of C1 per gratin of KBr. The solution was made accurately N/5 for precipitation reactions by volume, but was weighed and accordingly standard— ized with samples taken in the latter Way. These were precip- itated as AgBr . The following values were obtained, in grams of bromine, weight corrected to vacuum, per grafn in air of the solution :- A. B C D Average . Olš774O . Cl57731 O15777O . O15773O . Olš7743 Starch potassium iodide solution was used as described later to detect removal of the lºst traces of bromine. Schiff's fuchsive reagent was used for the same purpose When chloride was also present in the solution, as a test for bromine in the manner described by guerescº - The solutions were acidified with five normal Hºso, diluted from C. P. concentrated acid. On completion of heating, the esolutions were neutralized for the titration with a sodium * carbonate solution of equivalent strength, by titration, made - - - it: ſº from Kahlbaum's Nascog, ºur analyse. The periodate solution was used in the same manner as in the iodine determinations. To this was added the bromide solution, and acid as indicated in the various experiments, 73's then heated in the steam bath for 5 – 10 minutes. When the color of bromine no longer increased the air current was started. The vapors from the flask were drawn through a glass stoppered CaCl2 —U tube containing a test solution of starch and potassium - - iodide, for pure bromide solutions or Schiff's reagent for chloride mixtures. It was found essential for either test to cool the test solution by immersing the U – tube in cold water. The air current was drawn through the flask until no further test was obtained and then for about ten minutes longer. Preliminary experiments had indicated a fairly accurate determination of bromićie bromine in this manner, from solutions of about one normal H2SO4. However, it did not appear that HCl remained entirely unattacked in solutions of this acidity. Accordingly, the first experiments were made to determine the lowest acid concentration at which the bromide could be completely oxidized, within a reasonable time. The following determinations were all made in a total solution volume of 100 cc and 4.5 – 5 cc excess N/10 periodate. # Bromine 5\aso. Time of Bromine Error, mg. of taken grºs. added 3c. heating, min. found grºs. bromine. 1 .. 3619 2. 5 16O . 3530 -8.9 2 . 3598 5. O so .3528 –l. O 3 - 3602 5. O 29O. . 361C + .. 8 4 - 3596 1C.. O 45 . 3588 — . 8 5 , 359'? no.o 105 . 3592 – .. 5 6 : 3305 15. O 90 . 3602 – .. 3 7 . 3600 30. O 55 . 36CC O. C. 3608, + ... 1 3 : 33C7 2O. C. - 35 - 9 . .3599 35. C. 45 • 3 5 9 8 - 1. 74. 1.5 cc 5-Haso, should be deducted from the above, egºcent used in the reaction. The anomalous high result in #3 is easily explained by the - - long period of heating producing decomposition of the periodate. The next step was to determine whether a larger excess of the periodate would give complete oxidation with the lower acid contents. The following were made with 10 cc 5N-H2804 added or 8.5 cc as free acid, in 100 cc solution volume:- # Bromine N/10 As 203 Time of Bromine Error mg taken grºs. used. heating, min. found, grºs. of bromine 10 - 39.13 2.44 145 • 3793 –11. 7 ll. . 340l. 7. 48 90 .3395 — . 6 13 . 3038 - ll. 86 95 . 3040 + .. 3 13 . 256C 17.84 3C . 2563 + .. 3 14 . 2C-Cl 34. 34 55 , 2004 + .. 3 Three tests run at high dilutions were as follows :- Bromine Total 5N-H2SO4 Time N/10As 202 Bromine Error taken gins Vol. co - co, * heated cc. found. mg of * * min. - Ǻlò . bromine. 15 . 3638 3CO 3. 5 21C 4.65 . 36.18 —l. Q 16 . 3304 300 8.5 3?O 33. 10 , 216? -3. 7 l? . 2311 3CO l? ... O 150 33. 35 . 2311 O. O * Acid given as cubic centimeters of free 5N-H2804 per 100 cc of total volume. A number of determinations in the presence of chloride were now tried under the best conditions found in the work on pure bromide solutions. The separation is reasonably accurate in the presence of quite low concentration of chloride, but fails entirely as this is increased. Typical results obtained are as 75. shown below. In each case a total volume of 100 cc was used containing 8.5 cc 5N-H2894 as free acid. The completion of the liberation of bromine was tested by Schiff's fuchsine reagent. - * … T- - Căii. 3 Error # Bromine N/5 KCl Time of heat- N/10 As203 Br taken gras. C C . ing, min. CC a found gms mº bromine 18 .3632 5 105 4. 35 . 36.33 - . 1 19 . 36.33 5 100 4. 42 . 3628 — .. 5 20 .3838 5 130 4. 30 . 36.33 o.o 31 . 3631 10 130 4, 33 . 36.33 +l. 3 33 . 36.17 25 12O 4. 15 . 3648 +3. I It must be kept in mind in considering the results on the determination of bromine that the absolute errors in milligrams -- of bromine are not comparable to the corresponding figures for iodine. In the iodine determination .01 cc of N/10 arsenious acid is equivalent to .o.2 mg of iodine, in the bromine determination the same amount of reducing agent is equivalent to .08 mg of bromine. It is apparent, however, that the determination of brońine is not as satisfactory as that of iodine, because of the more tedious procedure required to bring the reaction to completion. The method recommended for the determination of bromine in the presence of iodine is as follows:– First, determine the iodine by the method aescribed. Dissolve another sample of the neutral salts in 5 – 10 cc of water, add an amount of Stands.rd periodate sufficient to oxidize the iodine to iodate, plus that estimated to be necessary to liberate bromine with 10 to 35 percent excess. Where the amount of iodine is high it will be necessary to introduce the sample in a capsule and heat to 980 C. before mixing, to avoid volatilization of iodine. Heat - tº - 76. the solution 10 – 12 minutes, then dilute to 100 cc and acidify with 30 cc of 5N-H2SO4 more than the amount necessary to liberate HBr from the bromide and l co for each 100 cc of basic periodate solution used. Heat and draw air through the solution until no further test for bromine evolution is obtained with starch iodide - t - ** *- - + * solution. Cool, neutralize with 20 cc 5M sodium carbonate, an t; it rate. In the presence of chloride, but not iodide, mix the perio— date in 10 – 25 % excess with the sample, at once. Dilute the solution until the concentration of chloridie does not exceed 35 mg. chlorine per 100 cc of solution. Add 5N-H280, in a mount equivalent to the bromide and periodate, as just described, and an excess of 8 – 10 cc per 100 cc of solution. Continue procedure from this point as is given above for bromide iodide mixtures, except use Schiff's fuchsine reagent as a test solution for bromine in place of starch iodide - 1 cc n/io Kºsºs solution - .007992 gm bromine. chlorine. Experiments were made tºo determine conditions for the quantitàtive evolution of chlorine from chlorides in the same manner as bromine. However, the acid concentration required was too high and the procedure necessary too long to warrant Well known methods. + consideration of this process in the face o There are probably no conditions under which such a method. º would have advantages over the ordinary silver titration. 77. RATE OF REACTION BETWEEN IODINE ANIJ PERIODIC ACID. The failure of the progress of the reaction of KI With K2H310s to conform to that expressed by the type equa- tions for any definite order, when formulated on the basis of the original reagents taken, and also the unexpected be- havior of this reaction in remaining unaffected by dilution, led to further inquiry into the mechanism of the process. One of the earliest experiments carried out in the study of the analytical method for iodine based on the oxidation of iodide iodine by periodates, was to let the reaction take place in acid solution. But it was quickly found to pro- ceed much more slowly in this way than with the neutral salts alone. On the other hand, Müller found that the reaction was retarded, also, by the presence of free alkali hydroxide. However, he calculated from electric potential measurements that the progress of the reaction in normal alkali, in the direction of 3 104 + I" --> 4 I03. involves a distinct diminution of free energy, He believed that a restraining influence existed which prevented the establishment of the stable equilibrium represented by the right hand members To overcome this, he resorted to the use of catalysers, and actually obtained an acceleration of the reaction by the presence of platinum black, as well as by neutralization of the alkali, 78. 70 - Bray ) took the view that oxidations of iodide to iodate, in general occur by the formation, first, of hypoiodite, and that all liberation of free iodine comes from the sec- ondary reaction of the hypoiodite with residual iodide. Accordingly, he formulated the reaction in question in this *8. Inn.61° e Tº t + t - ! t 3 (IO 4. I IO 3 * IOt) 3 IO ! :- I0'34 31' The quantitative relations of the results described in this paper, as well as a constant odor similar to that of iodoform, which was noticed continually in these experiments and can be ascribed to HIO, corroborate the assumption by Bray that hypoiodite is one of the intermediate products. However, Bray's equations seem to presuppose the presence of 1 mol. of Kl to each mole. of KIO4, or threefold the amount of KI required for the simple reaction 3 I0'4 + I t ---> 4. I0'3. The results already described with the reaction of the salt in equivalent proportions for the simple reaction throws little light on that occurring under the conditions considered by Bray. However, if the reaction takes place in two stages, there is absolutely no evidence to that effect given by the form of the curves obtained. In this work even the reaction between equiva- lent amounts of KT and K2H3106 was thought to be affected by too many undefined factoss to form the bases of theoretical 79. investigation, A fact either unknown to earlier investiga- tors of this reaction, or at least not considered by them, is the marked difference in the nature of the decomposition of KIO4, yielding ozone, from that of K2H3106 giving prac- tically no ozone. There is quite possibly a marked dis- tinction between the mechanism of oxidation realetions by solu- tions of these two salt, e. Yet as Pechard noted in 1898, one of the immediate products of the reaction of KI and K104 will be sufficient base to convert the residual KIO4 to K2H3106. Of course, this effect does not enter into the reaction between KT and K2H3106. However the latter components must also liberate a further amount of base in the formation of KIO3, and as Möller has shown, this in itself has a marked influence in retarding the reaction. For these reasons it was decided to study the reaction of iodine and H5106, thus eliminating at least such complica- tion as might arise from the presence of alkali cation. This was done by the following method: .0005 mol of iodine (for a 250 c.c. flask) weighed accurately, was introduced into the flask of the apparatus shown in the cut, and the latter immediately filled to about seven-eighths of its total capacity, with freshly boiled water, at o°C. Still at this temperature the flask was closed at the ground joint and attached to a high capacity mechanical vacuum pump, The stop-cock was opened cautiously and the system evacuated for 5-10 seconds, but closed again to the pump before any gas bubbles formed in the water and rose to ſ 80. the surface. The flas: Was then detached from the pump, quickly immersed in a large beaker of distilled Water and brought to 659–709 by heating the water surrounding it. Here it was rotated until the iodine crystals were entirely dissolved, then allowed to cool with the water outside to 50°C and transferred to a thermost at bath at 45°C. During the cooling from 65° to 45° it was essential to continue agitation of the solution in the flask to prevent the de- position of iodine on the inside of the neck, which might otherwise cool more rapidly th&n the solution. • OOi, 25 mol of H5I06 dissolved in about , 5 c.c. of water was introduced into the cup above the stop-cock. Then the stop-cock was opened slightly until this solution, and water at about 45° to rinse the sides of the cup, were drawn into the flask by the reduced pressure within. The solution within the flask was mixed thoroughly and the time noted. Finally enough more Water at 45° was brought into the flask in the same manner to bring the solution up to the graduation for 250 c.c. and the whole mixed thoroughly again. These last steps were all performed without admitting air, and without removing the flask from the thermostat bath, except for a few seconds at a time When shaken. After the color of the iodine had diminished considerably and the danger of loss by volatilization was not so great, usu- ally 6 to 8 hours after mixing the first sample was drawn, and thereafter at periods to obtain a well defined curve, These samples were drawn with a capacity pipette, this 81. rinsed into the titration flask, and finally a few cubic centimeters of 5% KT solution were shaken in the pipette to absorb any iodine vapors. These were added to the main solution. - The solution from the pipette was allowed to run directly into a cold, dilute, swirling, borax-boric acid mixture and instantly after the pipette had been emptied an excess of Kſ solution added. The iodine liberated was titrated with _N arsenious acid. - The **** of 45°C was selected for this work since the reaction at room temperature was so slow that several weeks and perhaps longer would have been required to approach completion. Also it was possible to obtain a higfooncentra- tion of iodine in the initial solution at the high temperature. The iodine used was prepared by the reaction of KIC3 with an excess of Kl in the presence of dilute phosphoric acid. The product of this ſeeaction was washed by decantation, thrown out on a centifuge in platinum cups, arted in vacuum over P295, and finally sublimed twice, the first time in presence of 3 05. The periodic acid was made by the electrolytic meth- od, was pure H5I06 by analysis, and gave no test for iodic acid. Three experiments were carried out following the pro- cedure given. The first, while giving similar results to the other two, was rendered inaccurate by the separation of a very small quantity of iodine in preparation. Accordingly, only the last twº experiments are considered. The final mixtures were, in each case, at the start 32. • CO2N with respect to free iodine, and .005 molar or . Ol normal for oxidation with respect to H5I06. 25 c. c. samples were drawn, which would initially have required 30 c.c. of N_ arsenious acid when titrated as described. 1C) The results of these experiments 2 and 3 were as follows: Titration Time # Minutes 118O 2890 4305 5700 6330. : Experiment 2. 100 As2O3= y: ( -x)K = # (-, … #y C C C. C. (30.00) (25.00) 14. 75 12. 29 9. 35 7, 8C 5. 87 4.93 3. 35 2,80 2, 90 3.42 0.0 0.0 Experiment 3. sº. As20326+ y (a-x)k 1OO O - (35.00) 23. 40 19, 50 9, 72 8, 1C 9. 40 7, 82 6, CO 5, CO 3. 85 3. 32 3. 3C 1.92 l, 60 1, 33 0.0 O. O. – 5 Kx = 3 (30-y) C C O.O l2, 71. 17, 20 20.08 22, 20 32, 58 (25.00) kx Titration Time Minutes O 645 3595 3695 4305 569C 71.45 857O O 5, 50 16, 90 17, 18 2O, CO 31.78 23. O3 23, 67 (35.00) 83. The calculated valued of the last two columns rep- resent respectively in cubic centimeteres of N solution of - 1OO oxidizing agent, the part of the original periodate solution remaining and that used in the reaction. As they arre, only by a constant factork = 2500 from the equivalent con- - centrations of periodic acid, they may be used in ratios for the latter. Further, as equivalent concentrations of period- ic acid and iodine were taken, these figures may be used in the same manner for concentrations of iodine. This assumes, as to the mechanism of the reaction, only that the net effect at any time corresponds to a progress of the process, 5 HIO4 + 2 I + H20 ---> 7 HIO3. Also, since equivalent amounts are used, the rate of the reaction may be expressed by some equation d: = K(e-x)", t where n is the number of molecules reacting and a-x is the equivalent concentration, common to all, of the residual amounts in solution. Calculations of the results obtained, according to various expressions for K derived from the integrated forms of the above equation with several values for n, revealed im– mediately that the reaction did not progress according to the formulation wherein n = 7, as indicated by the simplest chemi- cal equation. Nor did reaction appear to be of any high order. Instead, the results of the calculation of K for the first and second order were more nearly constant, as shown here . 84. Experiment 2 Experiment 3 Point, - -- - # lst order 2d order 1st order 2d order l –3– - l K=t in a-x K= x K=t ln a K: 3. t(a_x)a. 3-X t(a-x)= | - - l (603 X. 10–6) 351 x 10-7 386 x 156 175 × 10 7 2 403 305 434 32O, , 3 377 379 431 326 4. 383 556 373 373 5 370 590 361 - 475 6 360 674 7 342 832 The first titration of Experiment 2 gave a value which failed to accord with the other results obtained. The sense of this error was that the reaction had proceeded more rapidly than indicated by other values. However, the remainder of this same experiment was not affected. Accordingly, the devia- tion here is ascribed to a continuance of the reaction in the solution as being prepared for titration, occurring in the borax- boric acid solution before the KT solution was added. The technique of this operation had not been as well developed here 3,3 later. Titrations in which the residual periodic acid was still high would be more liable to such an effect. The calculated values of K plainly approach a constancy more nearly for the first order than any other . This condition means that the progress for the whole reaction is governed by some step wherein one of the reacting substances is present in relatively unchanged concentration throughout the process. 85, Such a substance could only be water, Hº or OHr. The most probable reaction of this type, because of the evident presence of HIO in the solution, was judged to be 19 + OH!---> H10. In the values of K for the first order, there remains the a systematic decrease as experiment continues. At first it was considered that this might arise from a slowing down of the reaction due to the following experimental error. As successive samples were drawn from the flask the gas space above the solution became larger, and it was possible that an amount of iodine had passed **, sufficient to affect appreciably the concentration of iodine in the solution. This possibility was examined by preparing several solutions in 100 c.c. flasks, but with correspondingly reduced amounts of the reagents to give the same concentrations as before, It was necessary to weigh; these exceedingly small - amounts of iodine by an assay balance • The periodic acid was taken as a concentrated solution. of known content by weight. These experiments were run simultaneously, but were opened successively through 4–5 days. The first solution was al- lowed to stay in the bath 14 hours after the first sample was taken and then another drawn and titrated, The results were as follows: 86. Solution Titration Time A. As293 K, 1st order - From start 1CO l 3, # # Minutes Ce Ce t = ln a-x l l 1691 14 - 50 424.5 × 10^* 2 3532 10.90 401 x 2 l 4040 6.84 367 x 3 l 6OOO 3,9C 340 x 2 8640 l.95 317 × The volatilization of iodine into the gas space plainly is not the cause of the decrease in the speed of the reaction, for the values of K decrease more regularly and more pronouncedly under the new conditions than under the old. Further, the second samples drawn from solutions l and 3, after standing, are in good accord in the series. An examination of the theoretical considerations affecting the rate of the reaction O - I + OH!---> HIO gives an insight into the difficulty. The concentration of of is not a constant but is ar- fected by two factors; first, the reduction of HIO4 to HIO3, if there is an appreciable difference in the degree of dissoci- ation of these acids at this dilution and ### temperature, will vary the H'concentration; second, the oxidation of I to HIO3 will certainly increase it. Either will change of concentra- tion by a reciprocal factor. Considering the net equation 5 HIO4 + 31 = 7. Hið3 87. and the concentrations at which the experiments were made we have the following (HIO4)o = .005 initial molar concentration of periodic acid. (I). = . . OC2 t? t? " of iodine, (H102) co = • CO7 final tº " " iodic acid. - - - º (H), - .005 x Y , where Y is the degree O YH10. Gio, of dissociation of the periodic acid. (H), - .005. Y. . a-x (.005 x + .003 x ).K., t %io, à + 3, 3. Kio, Assume that Éio. s Y - H 4. * = }^. This assumption 103 is discussed later, (#). = [.oos (a-x 4 x ) + .oogly t *** * Jr (1) - (.005 • OO3 2: ) . 3. Y - –l4 l (OH"). = 10 - 4.Y (2) t; •005+.003 x 3. (ſº), - .002 sex (3) Ta- dx O t Substituting (2) and (3) 14. 4.3 8A-X 10 ***. 4-Y : K (.co. 3, } •005 + .003 x } 3. # # Substitute a = .003 (for iodine) 14 - l * K, 10 • 4. Y | ... COE.E; 88. Let the constant expression -14 l l K. 10 24, Y 3. K2. Then { • OO6 + x } dt = K2 •003 - x) dix t = * {{#tº K X , --005 -- 003 | ſº • OO3 - x) dix = Keſ —I + l ln (.002-x) C O - K2 |-> - . OO'? in I.002-x) 4. • CO7 in.coal - *003 - Ks (.007 in Cô3-x - x) (.607 ºf:#-) - x) # K2 E. Recalculating the results of Experiments 3 and 3 to the new formula, the following results are obtained. Titration Experiment 3 Experiment 3 (331 x 10 °) 202 x 10" 235 × 10-8 252 x * 326 x tº 252 x * l 3 3 - 4. 339 x * 3.24.5 x " 5 333 x * 322 * th 6 7 - IHvA Hoºoºo ºogºv NNV ‘ºi'ſ as:Toogl *0.gſ ſe pțoſ oſ poſtea go Áq ! Aſą onpuoo retnoetoj， 008 89. The assumption that H5106 and HIO3 are equally dis- sociated under the conditions of the experiment is justified by the following considerations. The measurements of Ost- weidºns later of Rothmund and Drucks; as show that HIO3 in .007N solution at 25° is 97% dissociated while H510 s, if only the separation to H' and H41°s' is considered, is little more than 85% dissociated in .005 N solution. The amount of H' separated from the second hydrogen atom of HgIO6 may be entirely neglected, in these calculations, since even in solutions of Naroa, with no other H' present, it is not sufficient to color methyl orange. However, the conduct- ivity of periodic acid at 45° for successive dilutions W3.3 actually measured. The curve of molecular conductivity is given. Even for the initial concentration of these experi- ments, ºn H5Ios, this acid is dissociated 95%. Accordingly - 100 - the difference between the H' liberated from HEIO6 and that from HIO3 can effect a change in the results of but a few per cent which is within other experimental ©121°C1's e The effect of such a correction would be in the same direction as that already applied for the formation of HIO3 from I. The conclusion reached by these experiments is that the reactic.”hine and periodic cacid is governed by some inter- mediate reaction, in which the hydroxyl ioni is one of the re- acting substances. This intermediate step is supposed to be a reaction of free iodine with hydroxyl ion, since there is evidence of the presence of hypoiodous acid in the solution. 90. SUMMARY, In the experimental work described herein the following results have been attained, It has been shown that periodic acid will precipitate a number of cations quantitatively. Ferric periodate separates quantitatively from nitric acid solution, but the physical character of the precipitate prevents the use of the reaction for an analytical method. Potassium periodate is sufficiently insoluble in a solution containing 85% by volume of ethyl alco- hol and an excess of periodic acid to be removed from solution in this Way. This precipitate can either be weighed, or titrated for periodate oxygen to give a measure of the potassium present. When the titration is used this procedure constitutes a separa- tion from sodium, calcium, magnesium and sulfuric acid. Further, it has been shown that periodic acid or perio- dates react quantitatively in the oxidations of chromic salts to chromates, vanadyl salts to vanadic acid, iodides to iodates, and hydrobromic acid to bromine. The reaction with vana.dyl salts in acid solution is probably quantitative under conditions which allow no oxidation of chromic salts. But no end-point Could be devised to indicate the completion of the first reaction. Iodine can be determined in the presence of bromides and chlorides by oxidation of iodides to iodates with an excess of periodate in neutral solution, and titration of the residual periodate oxygen. Bromine can be determined similarly, alone or in the pres- ence of iodide or small amounts of chloride, after oxidation of 91. r in a current º he latt ilization of t at: e, and vol. e bromin --> tre HBr to of air . the tie of e is h iodin dic acid. Wit. perio reaction of º & "3, * -- roxyl ion is one of the reacting overned by a step in which hyd º -> In C36 e substa. 92. 7) 8) 9) REFERENCES * 3: # k >k ski: ; # * >k ºk ºk # F. Ammermuller and G. Magnus Ann, der Chemie ll:18 (1834) Bengie ser, Ann. der Chemie 17:354 (1836) C. Rammelsburg, Pogg. Ann. d. Phys, u chemie 44 :545, (1838) Langlois, Ann. chim. phys. (3) XXXIV:257 (1852 Ann. Ch. Pharm. LXXXIII:153 J. prak. Chem. LVI: 36 Rammelsburg and Marignac, Die neusten Forschungen in aer Krystallographiesaen Chemie, Liepzig, 1857 H. Kammerer, J. prak. Chem. XC:190 (1863) J. pharm. (3) XXXXV:106 C.G. Lautsch, J. prak. Chem. , C;65 (1867) Bull. soc. chim. (2) VIII:30 J. W. Ferniunds, J. prak. Chem. C: 99 Bull. soc. chim. (2) VIII:32 C. Rammelsburg, Berl, acad. Ber. , 1867; 691 J. prak. Chem. CIII, 378 io) c. Rammelsburg, Berl, acad. Ber, 1868, 207 11) C. Rammelsbury, Pogg. Ann. CXXXV, 493 12) c. Rammelsburg. Ber. a. deutsch, oh, Ges, 1869,17 Ber. d. deutsch ch. Ges. 1863, 70 J. prak. Chem. CIW, 434 loa). Pogg. Ann. 134; 368,499 - - 2 - Bull. goc. chim. (3) X: 357 12 a) Eer, deutsch, ch. Ges. 1869,147 93. 17) 18) 19) T J. Phillip, Pogg. Ann. CXXXVII: 319 Bull, soc. chim. (2) XII:350 K. Ihre , Ber. deutsch. ch. Ges, 1870, 316 14 a) C. Blomstrand, Ber, deutsch. ch. Ges-1870, 317 H. Kammerer Pogg. Ann. 138; 390 J. Thomsen, Ber deutsch. ch. Ges: 1872, 508 16 a.) Ber deutsch. ch. Ges: 1873, 6 A. Basarow, Ber, deutsch. ch. Ges, 1873 : 93 H. Kammerer, zeit, analyt. chem, 1873, 377 J. Thomsen, Ber. deutsch. ch. Ges; 1874;1498. 19 a) Ber. deutsch. ch. Ges. , 1874; 71 Fairley, Rep. Br. Assoc., 1875, 2 Ath. 48 Ogier, Compt. Rends 86, 733 Carnelly and W. Carleton-Williams,Chem. Soo. J. ; 37; lºš W. Ostwald, J. prak. Chem. (3) 32,300 C. W. Blomstrend. J prak. Chem. (3) 34, 433 34 a) J. prak. chem. (3) 40: 305 C. W. Kirºmins, Chem. soc. J. 51; 355 35 a) Rammelsburg, J. preſſ. chem. (2) 38,558–69 F. Selmons, Chem. Centra lö87, 503 P. Walden, zeit. phys. Chem. 2, 49 C. W. Kimmins, Chem. Soc. J. 55; 148 29 J. W. Retgers, Zeit. phys. Chem. 8; 23 C. W. Bloms trand, Zeit, anorg. Ch. , 1: lo–50 --- º E. H. Cook. Chem. Soc. J. 65:802–14 M. Hoehnel, Arch. d. Pharm. 332: 323 A.s, Eakle, zeit. Kryst. 26:558–588 B. Grützner, Archiv. der Pharm. 334:638 94. 35) 36) 37) 38) 39) 40) 41) 45) 46) 47) 48) Goochand Walker, Zeit. Anorg. Ch. 14:423–33 J. M. Crofts, Chem. Soc. J. 73:593–7 E. Pechard, Compt. Rend. 128:1101–04 37 a) Compt. Rend. 128, 1453 T- S. Tanatar, Ber. Deustch. ch. Ges. , 33:1103–5 A. Rosenheim and 0. Liebknecht, Lieb. Ann. 308:40–67 E. Pechard, compt. Rend. 130; 1705–8 Miolati and Mascetti, Gaz, chim. ital. 31, I, 93. Atti. Accad. Lincei (5) 14, 1:217 Arthur Michael and W. T. Conn. Am. Chem.Jas: 89–96 - º *- Hugh Marshall, Proc. Roy. Soc. Edinburgh 32: 388–90 tº Karl Garza rolli — Thurnlackh, Wien Acad. Ber. llC, II6 : 737–807 Monatschr. 23:955 (1901) Erich Müller. Zeit. Electrochem. 7: 509–16 A. Anstruc and H. Murco, Bull. soc. chim. (3) 37:939–30 Erich Müller, Ber. deutsch ch. Ges, 35:950–4 Erich Müller and 0. Friedberger, Ber. deutsch chem. Ges. 35:2858-9 A. O. Lamb, Am. Chem, J., 37:134-8 - F. Roques and A. Gerngross, J. Pharm. China (6) 16, 130–1 F. Giolitti, Gaz. chim. ital. 32, II, 340-54 Erich Muller, Zeit. Electrochem. 9:584–94 52 a.) Zeit. Electrochem. 9:707–16 W. Rothmund and K. Drucker, Zeit. Phys. Chem, 46:349 Erich Müller, Zeit. Electroch. 10:49–68 - 54 a) Erich Müller, zeit. Electroch. lo: º33-6 95, 56) 57) 58) 65) 66) 67) 68) 69) 70) 71) F. Giolitti, Atti. R. Accadi. dei Lincei Rena l:4, I: 217–20 S. R. Benedict, Am. Chem. J. 34: 531-5 O. Loewſ and K. Aso, Bull. of the College of Agriculture, Tokyo 7:1–6, Chem. Centralb. ??, 2:493 H. Brünner and R. Wellet, J. Prak. Chem. (2) 77 : 33 T. V. Barker, Chem. Soc. J. 93, I: 15–17 S. Bursarsky and Bela Horvath, Zeit. Ancrg. chem. 83:318-24 A. W. Browne and F. F. Sheterly. Am. Chem. J. 31:331 D. Vitali, Giorn. Farm. Chim, 59: 18–19 Lenz and Richter, Zeit. analyt. chem. ; 50; 537 v. Auger, Bull. Soo, chim. (4) 11:731-7 - E. Müller and W. Jacob, Zeit. anorg. chem; 8.3: 306–l4 E. Müller and G. wegian, Zeit. Analyt v. Chem. 53: 755–9 E. Corneo, Ann. Chim, et Phys (8) 29: 451–540 Rene Duprisay, Compt. Rend. 157: 1150–3 W. C. Brgy, Zeits. Phys. Chem. 54: 463 I. Guareschi, Zeit. Analyt. Chem, 53: 451 RULES COVERING USE OF MANUSCRIPT THESES IN THE UNIVERSITY OF MICHIGAN LIBRARY AND THE GRADUATE SCHOOL OFFICE Unpublished theses submitted for the doctor's degrees and deposited in the University of Michigan Library and in the Office of the Graduate School are open for inspection, but are to be used only with due regard to the rights of the authors. For this reason it is necessary to require that a manuscript thesis be read within the Library or the Office of the Graduate School. If the thesis is borrowed by another Library, the same rules should be observed by it. Bibliographical refer- ences may be noted, but passages may be copied only with the permission of the authors, and proper credit must be given in subsequent written or published work. Extensive copying or publication of the thesis in whole or in part must have the consent of the author as well as of the Dean of the Graduate School. This thesis by . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . has been used by the following persons, whose signatures attest their acceptance of the above restrictions. A Library which borrows this thesis for use by its readers is expected to secure the signature of each user. NAME AND ADDRESS DATE º, - º º º º - º - º º º º Fººººº. º º º Lºlº º ºº: º º ºº º º §§§ º º- º Tºº -º- - º º º º -- º - º º: º - T ºº - º º ºº:: º º º º º º º º º º º º ſº º º - º º - -- º ºº: º - º ºº: º º º º º º º T º - - - º º - - º º -º- -- º º sº º º º º º º - - º - º - T- - º º - º º º º º º º ºº º º º º -- º -º-º-º: º - ºrrºr º - º º º º º º º º º º - º º T º - º: º º - - º --- º: | º º - º sº - º º º - º º - º - - - º º - º - ºº: ſº- º - º º -º-º: º º ºº: º º --- º º º º -º-º: º º * º º-ºº: º º ºº:: -º-º: º º - º - º º º ºSººsº T º º º: ºº: ºriº º º - º º sº º ºl. - - º º - º º º - º º *** ***** - º - ºº:: º ºº:: º - - --- - º º º **º º - --- -º- º º-º-º-º-º-º-º-º: º º º sº sº ºsº.º. º - sº º º º º - º º º: º º º - º º ºº: º º º º º º º - º º º º º º º º: º -- ºº: - º º Sºº | º º- º º Tº º º º - º º º º - º, ºr -- - º º --- - ºº: º º º º º ºº: * º º º * sº º º-ºº: º * º T - - - - º º **. º - Fº º -º- º º º º --~~~~ ºr-º- - º º - --~~~~ º ---. - - º º ºº: --- º - tºº º: º: º - - º: º - º º º º º º º T * - * º º ºº:: º º - ºn.” º º º º - T º - º º - - --- --- º: º º º º *…*** - ºr -- nº ----- º º º º º - º º ºº:: º º- º - º --- º: º º º - - º º º - - - -- º -º-º: º - º º º º º º º - º º º ** **** ºº: **º- º T- ºil-ºr-º- rº- º - - º Tº - -º-º- --- - º | -º - º º - - º ----- -- - º - - - º º º - º-º: -º-º-º: - - º - º Tº - -- - --- --- - - - º ----------- T- ºº:: º º - Tº º |- - ---, ºr º - º º º º º º --- --- º º º º º -- ºº:: º º - tºº." … º º - --~~~~ diº. -- º: º º º- º º -- - º - -º- -- ". --~. T- - ----------, * - º - º - º - º *-* - º - --- rºtº- -º-º-º: º º º -º-º-º-º: º º ** º - - - - - º --- ------ -º-º: --- º º º º - º -º-º: --- º: º: * -- --- -atiºn º - º - º --- - º - ---. --- ºº:: T - --~~ º --~~~ º, - --- ---. -- - - - ººº- - - - - º º *…*.* -- - - - º - tº- º **** - - -------- --- º -*. - º'º. ** º -------- º -º-º: -- - º º ------- ºr º - **