A STUDY OF THE DECOMPOSITKON OF POTASSIUM FERRATE (V!) EN AQUEOUS $QLUTEON Thesls 50? ”we Dawn 0? pk. D. MECHIGéH STATE HNEVERSETY Richard Geoffrey Haire 1965 [Hams 6.2) .__ 3 1' Q4: _‘ -—‘ LI B RA R Y Michigan C x; Univcmm 4- Q P“ "'4'! "7 "7:. t V 4 Q ABSTRACT A STUDY OF THE DECOMPOSITION OF POTASSIUM FERRATE (VI) IN AQUEOUS SOLUTION by Richard Geoffrey Haire The decomposition of potassium ferrate (VI) in aqueous solutions pH 7.7 to 7 M in sodium hydroxide has been followed oxidimetrically, spectrophoto- metrically and by the volume of oxygen evolved in the decomposition. The ionic strength was maintained at two for buffers and sodium hydroxide solutions up to 2 M and maintained at approximately seven for 1 to 7 M sodium hydroxide solutions. The concentration of ferrate (VI) was varied from 4 x 10‘5 to 3 x 10‘2 M. A study of dilute ferrate (VI) solutions in which precipitation of Fe (III) did not occur, established that Fe (III) was important in the decomposition mechanism. The dependence of the decomposition rate on Fe (III) varied with the initial concentration of ferrate (VI): the concentra- tion of Fe (III) and the alkalinity of the solution. In solutions from pH 10 to 2 M in sodium hydroxide, the dependence of the decomposition rate on Fe (III) changed from first-order to one-half-order Richard Geoffrey Haire and to zero-order as the initial ferrate concentration was increased. The low solubility of Fe (III) in the pH 7.7 to 10 solutions did not permit a similar study for this pH range. For the more concentrated hydroxide solutions, the decomposition was inhibited by Fe (III) with the decomposition rate being first- order in Fe (III). With the higher ferrate concentrations used in the study, precipitation of an Fe (III) solid phase occurred during the decomposition. Decomposi- tion in solutions that initially contained a solid Fe (III) phase resulted in an increased decomposition rate for two hydroxide ion concentration ranges, namely pH 9 to 10 and 4 to 7 M sodium hydroxide with the rate being approximately twice as fast in the latter solutions. The decomposition rate in the pH 9 - 10 range was several times faster in solutions which contained a larger amount of solid Fe (III). 3 M ferrate (VI) The decomposition of 10- solutions was found to be second-order in ferrate and first-order in hydrogen ion between pH 7.7 to 9.5. A rate constant of 1.97 t 0.30 x 10'10 2- mole-‘2»min-l was found. The decomposition of <2 P?! "I (I. (T m Richard Geoffrey Haire 10-5 - 10-4 M ferrate solutions between pH 7.7 to 9.5 was first-order in ferrate (VI), but precipita- tion of Fe (III) changed the decomposition rate and only estimates of the rate constants were obtained. Values of 0.2 to 0.02 min"1 were found. The decomposition rate was also first-order in hydrogen ion for this ferrate concentration range. In solutions from pH 10 to 7 M in sodium hydroxide, the decomposition rate was first-order in ferrate and dilute ferrate (VI) solutions showed the above mentioned dependence on the Fe (III) concentration. For decompositions in solutions containing an Fe (III) solid phase, the Fe (III) concentration in the solution remained constant. The decomposition rate of ferrate (VI) was found to decrease from pH 7.7 to 9.5, then increase rapidly up to pH 12 and finally decreased in a uniform manner from pH 12 to 7 M sodium hydroxide. For ferrate (VI) solutions less than 3 x 10"4 M, the decomposition was inhibited by (0H)%. The decrease in the decomposition rate with hydroxide concentra- ‘2 M ferrate (VI) tion was greater for 10'3 - 10 solutions as a result of the inhibition by Fe (III). First-order rate constants in 1 to 7 M sodium Richard Geoffrey Haire hydroxide solutions ranged from 0.3 to 0.1 min"1 for 10"5 - 10-4 M ferrate (VI) solutions and 3 min"1 for solutions 10‘3 — 6 x 10‘3 to 2 x 10' 10'2 M on ferrate (VI). Mechanisms and rate expressions were suggested for the decomposition of ferrate (VI) in solutions with different compositions. The involvement of Fe (V) and Fe (IV) intermediates was postulated. The inability to identify intermediates or various iron species which may have existed in the decomposi- tion and the lack of information concerning the reactions of Fe (V) and Fe (IV) prohibited the development of a single mechanism or rate equation. Additional experiments included an attempt to prepare a pure compound containing iron in an intermediate oxidation state and an experiment to determine if isotopic exchange occurred between Fe (III) and ferrate (VI) in 7 M sodium hydroxide. Decomposition of high purity potassium ferrate (VI) in the solid state at elevated temperatures produced a mixed product, which decomposed during attempts to separate the components. The procedure for separating ferrate (VI) and Fe (III) in the exchange experiments was not satisfactory and the ferrate (VI) precipitates were contaminated with Fe (III). A STUDY OF THE DECOMPOSITION OF POTASSIUM FERRATE (VI) IN AQUEOUS SOLUTION BY Richard Geoffrey Haire A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1965 P" k p I < 2 {.5 Name Born Acad. Degre VITA Name: Richard Geoffrey Haire Born: May 18, 1935 in Chicago, Illinois Academic Career: Chicago Christian High School, Chicago, Illinois (1951 - 1953) University of Illinois, Navy Pier, Chicago, Illinois (1953 - 1955) North Central College, Naperville, Illinois (1958 - 1960) Michigan State University, East Lansing, Michigan (1960 - 1965) Degree Held: B. A. North Central College (1960) ii ACKNOWLEDGEMENTS The author wishes to express his gratitude to Dr. Andrew'Timnick for his guidance during the course of this investigation. The author also acknowledges the financial support of the National Science Foundation, the Socony-Mobil Oil Company and E. I. Du Pont De Nemours & Company. To Jack Holland, the author would like to extend his thanks for helpful suggestions and the loan of several instruments employed in this study. Special thanks go to the author's wife, Arlene, for her patience and encouragement during this investiga- tion and for the typing of this thesis. iii “I I ‘m-n ”TKJIZ J" 3') ‘ I . EMT" w“— EXPER TABLE OF CONTENTS VITA . . . . . . . . . . . . . . . s . . . . . ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . . . . . LIST OF TABLES . . . . . . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . . . . HISTORICAL . . . . . . . . . . . . . . . . . . EXPERIMENTAL . . . . . . . . . . . . . . . . . INSTRUMENTATION . . . . . . . . . . . . . CHEMICALS . . . . . . . . . . . . . . . . PREPARATION OF FERRATE (VI) . . . . . . . PREPARATION OF SOLUTIONS USED FOR THE KINETIC STUDIES . . . . . . . EXPERIMENTAL PROCEDURES . . . . . . . . . Kinetic Runs using Gas Buret Kinetic Runs using the Modified Chromite Method Kinetic Runs using the Direct Photometric Method Kinetic Runs using the Oxygen Analyzer RESULTS AND DISCUSSION . . . . . . . . . . . . General Rate Studies in pH 7.7 - 10.0 Range, 10‘3 M Ferrate (VI) Solutions Rate Studies in pH 10.3 to 2 M Sodium Hydroxide Solutions 10-3 M Ferrate (VI) Decomposition of Dilute Ferrate Solutions pH 7.7-2 M Sodium Hydroxide Solutions Decomposition of Dilute Ferrate Solutions, 1 - 7 M Sodium Hydroxide Solutions iv Page ii iii vi ix 13 14 14 16 17 18 18 21 24 28 29 30 32 82 124 Page Rate Studies in 1 - 7 M Sodium Hydroxide Solutions, 10"3 - 10'2 M Ferrate (VI) 142 Decomposition of Dilute Ferrate Solutions using a Galvanic Oxygen Analyzer 152 Decomposition of Solid Potassium Ferrate (VI) at Elevated Temperatures 154 Isotopic Exchange of Ferrate (VI) and Fe (III) 158 Discussion of the Previous Studies on Decomposition of Potassium Ferrate (VI) 161 SUMMARY AND CONCLUSION . . . . . . . . . . . . . 166 LITERATURE CITED 0 O O O O O O O O 0 0 6 O O O O 186 LImnv- V >1 0“. Page Rate Studies in l - 7 M Sodium Hydroxide Solutions, 10'3 - 10‘2 M Ferrate (VI) 142 Decomposition of Dilute Ferrate Solutions using a Galvanic Oxygen Analyzer 152 Decomposition of Solid Potassium Ferrate (VI) at Elevated Temperatures 154 Isotopic Exchange of Ferrate (VI) and Fe (III) 158 Discussion of the Previous Studies on Decomposition of Potassium Ferrate (VI) 161 SUMMARY AND CONCLUSION . . . . . . . . . . . . . 166 LITERATWE CITED 0 O O O O O O O 0 0 O O O C O O 186 F11 LIST OF FIGURES Figure 1. 2. 10. 11. Gas Buret Assembly for Oxygen Evolution . . Decomposition of Ferrate (VI) in Buffered Solutions pH 7.7 and 8.25 as Followed by the Modified Chromite Method . . . . . . . Decomposition of Ferrate (VI) in Buffered Solutions pH 8.65 and 9.10 as Followed by the Modified Chromite Method. . . . . . . Decomposition of Ferrate (VI) in Buffered Solution pH 9.10 as Followed by the Modified Chromite Method . . . . . . . . . Decomposition of Ferrate (VI) in Buffered Solution pH 9.10 Initially Containing Fe (III) - Decomposition as Followed by the Modified Chromite Method . . . . . . . . . Decomposition of Ferrate (VI) in Buffered Solution pH 9.4 as Followed by the Modified Chromite Method . . . . . . . . . . . . . Decomposition of Ferrate (VI) at 25° in Buffered Solution pH 9.2, Ionic Strength 0.1, as Followed by the Modified Chromite Method . . . . . . . . . . . . . . . . . . Decomposition of Ferrate (VI) at 40° in Buffered Solution pH 9.1, Ionic Strength 0.1 as Followed by the Modified Chromite Method . . . . . . . . . . . . . . . . . . Decomposition of Ferrate (VI) in Buffer Solution pH 7.70 and 8.25 as Followed by the Oxygen Evolution Method . . . . . . . Variation of kexp,‘With pH of the Solution for pH 7.7 - 9.5. . . . . . . . . . . . . Decomposition of Ferrate (VI) in Buffers pH 10.3, 10.6 and 11.1 as Followed by the Modified Chromite Method . . . . . . . . vi Page 19 33 34 37 38 39 42 43 45 49 6O Fig 17. 18. 19. 21. 22. 23. Figure Page 12. Effect of the Ionic Strength on the Decomposition Rate of Ferrate (VI) as Followed by the Modified Chromite Method . 63 13. Decomposition of 4 x 10.3 M Ferrate (VI) in Sodium Hydroxide Solutions as Followed by the Modified Chromite Method . . . . . 64 14. Decomposition of 3 x 10"4 M Ferrate (VI) in pH 11.1 and 11.9 Solutions as Followed by the Oxygen Evolution Method . . . . . . 66 15. Decomposition of Ferrate (VI) in pH 12.6 Solution as Followed by the Oxygen Evolution Method - Effect of Presaturating the Solution with Oxygen . . . . . . . . . 67 16. Decomposition of Ferrate (VI) in 1.3 M Sodium Hydroxide Solution as Followed by the Oxygen Evolution Method . . . . . . . 69 17. Variation of kexp. with pH of the Solution for pH 10 - 14 using the Modified Chromite Method . . . . . . . . . . . . . . . . . . 71 18. Variation of kexp. as Function of Sodium Hydroxide Molarity . . . . . . . . . . . . 73 19. Decomposition of Ferrate (VI) in pH 7.70 and 8.25 Borate Buffers as Followed by the Modified Chromite Method . . . . . . . . . 85 20. Decomposition of Ferrate (VI) in pH 8.25, 8.65 and 8.40 Carbonate Solutions as Followed by the Direct Photometric Method . 87 21. Variation of kexp, for pH 7.7 - 9.5 Range for Dilute Ferrate (VI) Solutions . . . . . 89 22. Decomposition of Ferrate (VI) in pH 10.0 Phosphate Buffer as Followed by the Direct Photometric Method . . . . . . . . . 93 23. Decomposition of Ferrate (VI) in pH 10.6 PhosPhate Buffer as Followed by the Direct Photometric Method . . . . . . . . . . . . 94 vii r0 2 27. 28. 29. 30. Figure Page 24. Effect of Fe (III) on the Decomposition Rate of Ferrate (VI) in pH 10.6 Phosphate Buffer as Followed by the Direct Photo- metric Method . . . . . . . . . . . . . . 97 25. Decomposition of Ferrate (VI) in pH 10.3 PhOSphate Buffer as Followed by the Direct Photometric Method - k'exp. plot. . 103 26. Decomposition of Ferrate (VI) in pH 10.3 Phosphate Buffer as Followed by the Direct Photometric Method - kuexp. plot . . . . . 104 27. Decomposition of Ferrate (VI) in pH 13.4 Sodium Hydroxide Solution as Followed by the Direct Photometric Method . . . . . . 114 28. Decomposition of Ferrate (VI) in 7 M Sodium Hydroxide Solution as Followed by the Direct Photometric Method . . . . . 128 29. Decomposition of Successive Portions of Ferrate in 7 M Sodium Hydroxide salutions O O O O O O O O O O O O O O O O 129 30. Decomposition of Ferrate (VI) in 7 M Sodium Hydroxide as Followed by the Direct Photometric Method - k"gxp. plot . 130 viii Table vi 7L ECO .NL ES~C~L VII. 21*; VIII. EEC nonupc Table II. III. IV. VI. VII. VIII. LIST OF TABLES Page Decomposition of Ferrate (VI) as Followed by the Modified Chromite Method, pH 7.7 to 10. . . . . . . . . . . . . . . . . . . 51 Decomposition of Ferrate (VI) as Followed by Oxygen Evolution, pH 7.7 to 9.5 . . . . 55 Decomposition of Ferrate (VI) as Followed by the Modified Chromite Method, pH 10.3 to 14° C O O O O O O O O O O O O O O O O O 74 Decomposition of Ferrate (VI) as Followed by the Oxygen Evolution Method, pH 10.3 to 14 O O O O O O O O O O O O C O O O O O I O 78 Decomposition of Dilute Ferrate (VI) SOIUtiODS, PH 797 - 905 o o o o o o o o o 91 Decomposition of Dilute Ferrate (VI) Solutions, pH 10 - 13 as Followed by the Modified Chromite Method . . . . . . . . . 106 Decomposition of Dilute Ferrate (VI) Solutions, pH 13 — 2 M Sodiumnfiydroxide as Followed by the Direct Photometric methOd O O O O 0 O O O O O O O O O O O O 118 Decomposition of Dilute Ferrate Solutions, 1 - 7 M in Sodium Hydroxide as Followed by the Direct Photometric Method . . . . . . 135 Decomposition of Ferrate in 1 - 7 M Sodium Hydroxide as Followed by the Modified Chromite Method . . . . . . . . . . . . . 148 Decomposition of Ferrate in 1 - 7 M Sodium Hydroxide as Followed by the Oxygen Evolution Method . . . . . . . . . . . . . 150 ix INTRODUCT ION nn'I-‘Y '. - F90, oxid when Ferrate (VI) compounds are substances containing FeO42” ion in which iron exhibits its maximum known oxidation state. Solid ferrate salts are stable when protected from moisture and heat. Aqueous solutions of soluble ferrate salts are wine red to deep purple in color and the solutions decompose with widely varying rates, depending on the temperature and composition of the solutions. Immediate decomposition occurs when these salts are added to acidic solutions, whereas addition to a concentrated alkaline soluflon at lower temperatures produces a ferrate solution which exists for several days. The potassium salt shows no tendency to dissolve in solvents except aqueous solutions or solvents which it oxidizes rapidly. Rapid decomposition occurs if an oxidizible, material is present and aqueous solutions of dioxane or ethyl alcohol decompose ferrate (VI). Ferrate (VI) ion in general is a stronger oxidant than permanganate and can oxidize aqueous ammonia to nitrogen. In the absence of an oxidizable material, ferrate (VI) solutions decompose according to the following equation shown for the potassium salt: 2K2Fe04 + (2+x) H20—) 3/2 02 + Fe203oxH20 + 4KOH The barium salt is insoluble in aqueous solutions that are not acidic and decomposition occurs very slowly. Sol ult amo: mg for: Solid ferrate (VI) compounds decompose when heated, ultimately forming oxygen and iron (III) compounds among the reaction products. The kinetics of aqueous decomposition may be followed by several methods. Since ferrate (VI) is 'a colored species, the absorbence of the solutions may be followed at 505 mp. The decomposition may also be followed oxidimetrically, making use of the oxidiz- ing power of ferrate (VI). Measurement of the oxygen formed in the reaction may also be employed. Following the decomposition by measurement of the hydroxide ion formed is not practical. The decomposition of aqueous solutions has been discussed both qualitatively and quantitatively by several investigators. The most recent work is that of Magee who correlated ferrate decomposition with the change in oxygen pressure. The present work was started prior to the appearance of Magee°s work and was continued in an attempt to resolve the questions pertaining to the decomposition process of ferrate raised by the other investigators. This work extends the ferrate concentration ranges studied and compares the different methods of following the decomposition of ferrate, including the previously untried method of following the oxygen evolved volumetrically. .- 'I . " .. ' ' orarl The effects of factors such as hydroxide ion concentration, pH, ionic strength, stirring, presence of products, oxygen and Fe (III), and temperature on the decomposition process are studied. HISTORICAL Th pertain1 early re resultin preparat (l), in on diss: fusion c in varic alkaline electro] iron as are ques CODClusJ is the ( formed ferrate GmP (3: a“ 0X1d« review 1 Prior t: T] COmPOun( SchreyQ] There are many references in the literature pertaining to ferrate (VI) compounds. In most of the early references, the appearance of the ferrate resulting from various chemical reactions and the preparation of ferrate salts are discussed. Stahl (l), in 1702 mentioned a purple solution resulted on dissolving the residue from a potassium nitrate fusion of iron filings. Ferrate has also been formed in various alkaline fusions containing iron, by alkaline chlorine oxidations of Fe (III) and during electrolysis of potassium hydroxide solutions using iron as the anode. Some of these early preparations are questionable and there exist discrepancies in conclusions that have been drawn. An example of this is the claimed preparation of silver ferrate (VI) formed in the reaction between silver nitrate and a ferrate solution to give a black precipitate (2). Gump (3)has shown that the precipitate was actually an oxide of silver and not a ferrate salt. A good review of the early literature of ferrate chemistry prior to 1952 is given by Gump. The preparation of a high purity ferrate (VI) compound was first reported in 1951 by Thompson, Schreyer and Ockerman (4). Their procedure involved a sodium hypochlorite oxidation of iron (III) nitrate in a concentrated sodium hydroxide solution, followed by precipitation of the potassium salt from a con- centrated potassium hydroxide solution. With additional purification, it is possible to obtain a purity of 98 - 99%. The above method was used by this author for the present investigation. Gump (3) has prepared several ferrate salts by using a similar procedure of separating the desired ferrate salt from a concentrated metal hydroxide solution. Although none of the salts were obtained as pure as the potassium salt, Gump did prepare the alkali metal ferrates and some of the alkaline earth ferrates. The difficulty of obtaining pure ferrate salts other than the potassium salt resulted from the higher’ solubility of the lighter metal salts and the unavail- ability of concentrated solutions of the heavier metal hydroxides. Gump also reported the preparation of the cadmium, zinc and lanthanum ferrate salts in addition to the above salts. Krebs (5) has shown that the structure of ferrate is similar to chromate. Both are tetrahedral solids with approximately the same cell dimensions. Others (6) have compared potassium ferrate and chromate by x-ray studies and Jellinek (7) lists the cell dnmensions for barium ferrate (VI). Symons 23 31. (8,9) have studied the electron spin resonance and electronic spectra of ferrate. Schreyer (10) Obtained a qualitative spectral transmittance curve for potassium ferrate. Kaufman and Schreyer (11) reported a quantitative study of the absorbance of ferrate solutions from 320 to 1000 mp. They reported a maximum absorbance at 500 mp and found a linear relationship between absorbance and concentrations of ferrate. A molar absorptivity of 1.13 x 103 is reported. Wood (12) also determined the molar absorptivity of ferrate solutions and arrived at a value of 1070:: 30 1-mo1e-l-cm-l. Kochanny and Timnick (13) investi- gated the spectra of ferrate formed when hydrogen peroxide is added to an alkaline solution of disodium ethylenediamine tetracetate that contains a small amount of ferric hydroxide. It is of interest to note that addition of hydrogen peroxide to a ferrate solution normally results in an immediate decomposition of the ferrate (14). Kochanny and Timnick found that EDTA was necessary to form the ferrate in the peroxide system and that the ferrate absorbance peak is shifted to longer wave lengths when EDTA is present. Carrington and Schonland (15) have also investigated the ferrate spectra and have compared it to the spectra of other oxyanions that contain one or two unpaired electrons. of f: by W: stand Fe(OH to be ferra speciq discus have q soluti 96nera. Essent; and alk and the more re; conclude deCOmPOs faCtOr O that figue p0taSSium stability‘ EffECt Of The heat of formation, entropy and the free energy of formation of aqueous ferrate ion have been determined by Wood (12). The values were of -11513_1 Kcal/mole, 9 j; 4 e.u. and—771t2 Kcal/mole respectively. The standard electromotive force of the half reaction, Fe(OH)3-+ 5 OH'—9»Fe04 = + 4H20 + 3e, was estimated to be -0.72 i 0.03 volts. ‘Wood suggests that at pH 10, ferrate decomposes slowly and ferrate is a dinegative species at this pH. The stability of aqueous solutions has been discussed by several authors. Gump et 31. (3,16,17) have qualitatively discussed the stability of ferrate solutions. Schreyer and Ockerman (18) have also generalized on the stability of ferrate solutions. Essentially these authors conclude that the temperature and alkalinity of the solutions are important factors and that concentrated solutions appear to decompose more rapidly than dilute solutions. Gump gt 31. conclude that light appears to have no effect on the decomposition rate and that stirring is an important factor on the rate of decomposition. Gump (3) found that aqueous solutions of cesium, rubidium and potassium ferrates had approximately the same stability. Schreyer and Ockerman have studied the effect of various salts on the rate of decomposition and state that potassium chloride or bromide, the nitrate, carb reta and x comp: F8201 U follc 10 carbonate or chlorate of potassium or sodium chlorate retard aqueous decomposition. Calcium, strontium and magnesium salts, oxides, peroxides and organic compounds accelerate the decompositions. Hydrous Fe203 caused rapid decomposition. These authors followed the pH of aqueous ferrate solutions as a function of time. They concluded from the study that more concentrated solutions decomposed completely before attaining a constant pH level while dilute solutions attained a pH level which persisted for a short time before reaching the final pH value. One of the earlier quanitative decomposition studies was by Hinsvark (l4). Decomposition of alkaline solutions of potassium ferrate was followed spectrophotometrically at 30°. Studies were made in 3 4 4-9N sodium hydroxide solutions with 10‘ -10' M ferrate. The pseudo-first-order rate constants ranged from 0.10 to 0.029 min"1 for 4 and 9N sodium hydroxide respectively. A mechanism involving hydroxyl radicals was suggested. Jezowska-Trzebiatowska and Wronska (19) studied the decomposition of potassium ferrate in 7-10 M potassium hydroxide. Ferrate concentration ranged from 6.5 x 10'"4 to 2.5 x 10.3 M solutions and the decomposition was followed oxidimetrically. Studies were made at 200 and 30° and the decomposition rate was found to be first- 4 order ions Wrons potas conce The d1 Wrons} ferrat when a are mo than w sugges' neutra. was in? Wronska and POt to dete 11 order in ferrate. Inhibition was noticed with sodium ions while rubidium ions accelerated decomposition. Wronska (20)studied the decomposition of aqueous potassium ferrate, at 20° and 300 with ferrate 3 to 1 x 10'2 molar. concentrations of 2.5 x 10- The decomposition was again followed oxidimetrically. Wronska noted that in general, dilute solutions of ferrate are more stable when initially neutral than when alkaline and the higher ferrate concentrations are more stable in concentrated alkaline solutions than when the solution is initially neutral. Wronska suggested that ferrate decomposition in initially neutral solutions was second-order in ferrate and was inhibited by hydroxide ion formed in the reation. Wronska (21) also looked at the decomposition in sodium and potassium dibasic phosphate solutions in an attempt to determine the influence of the ionic environment on the decomposition rate. Magee (22) studied the decomposition of ferrate solution at 25° in buffers of pH 9.2, 9.6, 10.0, 10.8, 11.1 and sodium hydroxide solutions of pH 12.2, 13.0, 13.8 and 4.9, 7.1 molal. Decomposition was followed by measuring the oxygen pressure above ferrate solutions. The decomposition rate was found to be second-order in ferrate at pH 9.2 and 9.6 and first-order in the other solutions. The dependence of the rate on pH of the solution was determined. T of Fe ( Bunsen Klemm (. The Na4£ hydroxic barium 5 solution one hour K3FeO4 a: to form E oxidation aqueous s 12 The preparation and some properties of compounds of Fe (IV) and Fe (V) have been reported. Scholden, Bunsen and Zeiss (23) prepared Na4FeO4 and Scholden and Klemm (24) prepared barium and strontium Fe (IV) compounds. The Na4FeO4 disproportionates immediately in dilute sodium hydroxide solutions to form Fe (III) and Fe (VI). The barium salt, BazFeO4, decomposes more slowly in aqueous solutions complete decomposition taking approximately one hour. Klemm and Wahl (25) reported the preparation of K3Fe04 and found that Fe (V) disproportionates in water to form Fe (III) and Fe (VI). The existence of these oxidation states and knowledge of their reactions in aqueous solutions are important for understanding Ferrate(VI) decomposition in solution. EXPERIMENTAL 14 W The following instruments and equipment were employed to make the appropriate measurements: Beckman expanded scale pH meter with a Beckman saturated calomel - Beckman E—2 type glass electrode pair: Beckman Model DB Spectrophotometer, cell compartment maintained at 400C 1.0.10 for kinetic studies; Cary Model 14 Recording Spectrophotometer: Precision Scientific Galvanic Oxygen Analyzer: Sargent SRL Recorder used in conjunction with the DB Spectrophotometer and the oxygen analyzer. CHEMICALS The following chemicals were used without further purification, unless otherwise stated, for preparation of reagent solutions described in this study. Barium Hydroxide Baker's Analyzed Reagent Boric Acid Reagent Grade Allied Chemical Company Chromium Chloride, Reagent Grade Hexahydrate Allied Chemical Company l,S-Diphenylcarbohydrazide Eastman Kodak's White Label Iron - 59-P tracer Oak Ridge National (enriched) as FeC13 Laboratory solution Iron wire Analytical Reagent Mallinckrodt Chemical Works 15 Magnesium Perchlorate, Anhydrous Nitrogen Oxygen Reagent G. Frederick Smith Company Prepurified Matheson Company Industrial, 99.9% General Dynamics Corporation The oxygen was passed through a column of sodium hydroxide, then through concentrated H2804 and finally through a tube of ascarite-anhydrous magnesium perchlorate before use. ' Potassium Hydroxide Sodium Bicarbonate Sodium Hydroxide Pellets Baker's Analyzed Reagent J. T. Baker Chemical Company Reagent Grade Fischer Scientific Company Baker's Analyzed Reagent J. T. Baker Chemical Company Fresh pellets were washed quickly with water and a saturated solution made and stored in a polyethylene bottle. After several days standing, one-half of the clear saturated solution was removed to another polyetheylene bottle and stock solutions were made by dilution of this solution. Sodium Perchlorate, Monohydrate Reagent Grade G. Frederick Smith Company The reagent was recrystallized twice from water. An approximately 8 molar stock solution was made and was standardized gravimetrically by weighing the tetraphenylarsonium perchlorate salt. The weighing of ignited residues from aliquots of the stock solution gave identical analytical results. Stock solutions were made from this standard solution. Dibasic Sodium Phosphate, Heptahydrate Reagent Grade Allied Chemical Company 16 Tetraphenylarsonium Reagent Grade Chloride G. Frederick Smith Company Distilled water All solutions used in the kinetic studies were made using distilled water processed as follows: Laboratory distilled water was passed through a de-ionizing resin mixture and then re-distilled from alkaline permanganate. Preparation of Potassium Ferrate (VI) The solid potassium ferrate (VI) was synthesized by the method of Schreyer, Thompson and Ockerman (26), and was analyzed by the chromite procedure described ‘with the synthesis. This synthesis along with other methods of purification and analysis are reported (27,28,29,30,3l), but the above chromite procedure was found satisfactory. The chromite procedure involves oxidation of Cr (III) to Cr (VI) in saturated sodium hydroxide, followed by acidification with sulfuric acid and titration with standard ferrous solution to the diphenylamine sulfonate end point. After several recrystallizations, the potassium ferrate, which was prepared, obtained a purity of 98.1;t 0.1%. This material was used until the purity dropped to 95%. Samples of 95% purity were used in experiments in which Fe (III) was added or in which considerable decomposition of K2Fe04 resulted during the dissolution of this compound. were stoc to v were cont: bottJ hydro 17 Preparation of Solutions Used for the Kinetic Studies The sodium hydroxide-sodium perchlorate solutions were made by adding the appropriate amounts of the respective stock solutions to 50 m1 volumetric flasks and diluting to volume at 20°C. Aliquots of the resulting solution were titrated to establish definitely the sodium hydroxide content. The solutions were stored in polyethylene bottles which had been conditioned with 10M sodium hydroxide. .The various buffer solutions were made by dissolv- ing the appropriate reagent to make a 0.1 molar solution and then adding the appropriate amount of standard sodium hydroxide to give the approximate pH value desired. Borate, carbonate and phosphate buffers were employed. Sodium perchlorate was then added to give a total ionic strength of two. The pH values of the buffers and the dilute sodium hydroxide solutions were measured. Corrections for sodium ion content were applied as given in the monograph supplied by Beckman with the E—2 glass electrode. For 40°, measurements with 2 molar sodium ion concentration, corrections of 0.1 pH unit or less were applied up to pH 12.3. For 11M sodium hydroxide, the correction was 0.40 pH unit. For 0.0580 and 0.945 molar sodium hydroxide, excellent agreement was found between experimentally measured pH and pH calculated using activity coefficients reported by Kegeles (32). dec: by n solt Comp volu empl: by a deliv Figur 7.00m such 1 be lou at 40C extend mercur: m39n6t1 inside mounte 120 rpm Tr potassiu attachin. with the 18 Experimental Procedures Kineticfpgns using a gas buret: Kinetic data for decomposition of aqueous ferrate solutions were obtained by measuring the rate of oxygen evolution from the ferrate solutions with a gas buret—manometer (Arthur Thomas Company). The buret had a maximum volume of 7m1 and volumes could be estimated to 0.002m1. Mercury was employed as the manometer fluid. The buret was attached by a ball and socket joint to a calibrated reservoir- delivery system and reaction flask of the author's design, Figure l. The solution reservoir had a capacity of 7.00ml. The system was mounted on an aluminum frame, such that the entire system including the buret, could be lowered into a large thermostated water bath kept at 40°C 1 .050. An extension of the leveling bulb extended above the bath and allowed adjustment of the mercury levels. Stirring was provided by a glass magnetic stirring bar (0.6cm x 0.6cm diameter) located inside the reaction flask and was driven by a magnet mounted in the bath. A stirring rate of approximately 120 rpm was maintained. 'The procedure consisted of weighing the solid potassium ferrate into the dry reaction flask and attaching the flask to the delivery system filled with the desired solution, which had been presaturated 19 FIGURE 1. Gas Buret Assembly for Oxygen Evolution 3 - way stopcocks {)4 Filling Port Solution 4 Reservoir g. Teflon Reaction Flask h Magnetic Bars M V Nylon Support {—— Buret IL: 20 with oxygen. The system was evacuated, filled with dried oxygen and the reaction flask sealed off from the system. The flask-reservoir system was attached to the gas buret and the unit immersed in the bath. After temperature equilibrium had been attained, the pressure that had built up in the system was released and the solution was then mixed with the solid ferrate. Volume readings were periodically taken after adjusting the closed system to atmospheric pressure. Volume corrections were made if the atmospheric pressure changed during the run. The initial volume, Vo, was recorded a few seconds after adding the reaction medium to the solid ferrate. When no further volume change occurred on standing, the final volume, Va), was recorded. Graphs were made by plotting H2 or log (V00 - V) versus time for so _ second and first-order reactions respectively, where V is the volume at time, t. For second-order plots, the above gives a more uniform representation of data and requires only conversion of (Wm) - V0) to the correspond- ing ferrate concentration for calculation of the experimental rate constant. The second-order experimental rate constant was obtained from the relationship, Slope = (Von - Vo) kexp.'where (Wu) - Vo) represents the molarity of 21 ferrate (VI). (Wm: - Vo) was calculated for 400 from the ideal gas law as follows: (Va. - v.) = 4/3 x (-"—;%(-)L’ X Vol. 02 0.0820 x 313.l° K x 7.00 where V.P. vapor pressure of the solution, 7.00 = m1 of solution delivered, 4/3 = stoichiometric ratio, in moles of ferrate per mole of oxygen produced, and the volume of 02 is in ml. The vapor pressures for the different solutions were estimated from values listed for different salt solutions (33). For first-order plots, the experimental rate constant was obtained by dividing the slope of the straight line in the plot by -2.303. Kinetic Runs Usingithe Modified Chgomite Method The titrimetric chromite method of Schreyer, Thompson and Ockerman (26) for analysis of ferrate can be used to follow the rate of decomposition of relatively large concentrations of ferrate. Although smaller burets and a more dilute titrant could be used, it is desired to extend this method so that 10"4 - 1073M solutions could be easily handled. This was accomplished by modifying the chromite method so that the chromate formed was determined spectrophoto- met nat aPP‘ forr ferr to a sodi' watel final using aside chrom. 0.25 1 detern using SPeCie; conver1 P°r COn as Aofl. and 109 slope o f 22 metrically using 1, 5-diphenylcarbohydrazide to form a highly colored species with chromate (34). The exact nature of the colored species is not known, but it appears to be a Cr (III)-diphenycarbazone species that forms only with Cr (II) or chromate. The procedure consisted of taking aliquots of the ferrate solution being studied and adding each aliquot to a volumetric flask containing Cr (III) in saturated sodium hydroxide. The solution was then diluted with water and acidified with 6M sulfuric acid to make the final solution approximately 0.2N in acid (pH 0.8, using pH 0-l.5 hydrion paper). The solutions were set aside until the kinetic run was complete and the colored chrominum species were formed by adding a quantity of 0.25 percent 1, 5-diphenylcarbohydrazide solution (50% acetone-water). The absorbance of the sample was determined at 540 mp with a beckman DB Spectrophotometer using 1cm cells. The absorbance of the colored chrominum species could be used directly for plotting or could be converted into the corresponding ferrate concentration. For convenience and uniformity, the data were plotted as Ao/fi values versus time for second-order reactions and log A/Ao for the first-order reactions. The second- order experimental rate constant was obtained from the slape of the straight line using the relationship, 51: co: ca] cho was fro: for enou depe Cr (1 hydra satur place. dipher and 1C ferrat Propor a larg. vOlumet Were ta AlthOUgj by titre “Our: ts 23 slope, 3 Aok, where Ao was converted into the correSponding ferrate concentration by using a calibration chart. The first experimental value was chosen for A0. The first-order experimental rate constant was obtained by dividing the slope of the straight line from the log plot by -2.303. This modified chromite method was applicable for very dilute solutions to solutions concentrated enough for titration. Three variations were used depending on the ferrate concentration. For very dilute solutions, 1 ml of an aqueous Cr (III) solution (300 grams Cr (III) chloride hexa- hydrate in 250 m1 of water) was added to 40 m1 of saturated sodium hydroxide and 2 m1 of this solution placed in 25 ml volumetric flask. One ml of the diphenylcarbohydrazide solution was used. Fifty m1 and 100 m1 flasks were used for more concentrated ferrate solutions and the reagents were increased proportionately. For concentrated ferrate solutions, a larger amount of Cr (III) was used in a 50 m1 volumetric flask and aliquots from each 50 ml flask were taken for color development in 25 m1 flasks. Although this concentration range could be determined by titration, there is the advantage that much smaller amounts of ferrate are required for the modified 24 chromite method. The concentration range for color development was 2 to 20 micrograms of chromium in a 25 m1 flask. Reasonably low blanks were obtained for the procedure. Standard solutions were prepared from 0.1N potassium dichromate solutions and calibration charts prepared from these standards. Addition of Cr (III) to standard solutions prior to color development did not effect the net absorbance. The method did not appear to be sensitive to small variations in procedure. Kinetic runs were made by weighing solid potassium ferrate into a dry 50 ml Erlenmeyer flask and pipeting the desired solution into a second flask. The solution was presaturated with oxygen and both flasks placed in a thermostated water bath maintained at 40;: .OSOC. After temperature equilibrium had been attained, the solution was transferred to the sample flask. A magnetic stirring bar (2.5 cm x 0.6 cm) was used for stirring. Samples were then periodically withdrawn for the modified chromite method of analysis. Kinetic Runs Using phe Digect Photometgic Method A The cell compartment of the DB Spectrophotometer wasthermostated for 400 i 0.l°C and special rectangular glass cells with a path length of 2.70 cm were used. For the kinetic studies, the instrument was set at a wavelength of 505 my and the absorbance of ferrate (VI) was r minut of th conce absor The p Solut to re of th amoun absor This alkal than for s Cloud: hYdrO: Fe (I: 25 was recorded with a chart speed of one inch per minute. A very simple approach was used for the preparation of the ferrate solutions. Since trial runs indicated that the decomposition was first-order in ferrate, the rate constant should not be dependent on the concentration. Hence, it was not necessary to know the concentration and by merely adding small amounts of solid potassium ferrate to the solution, the rate of decomposition could be followed. Actually, the concentration could easily be calculated from the absorbance data and the molar absorptivity of ferrate. The procedure consisted of placing the cell with the solution into the compartment and allowing the solution to reach the desired temperature. The initial absorbance of the solution was recorded as a blank and then a small amount of K2FeO4 was mixed directly in the cell. The absorbance was recorded until decomposition was complete. This procedure worked well for the more concentrated alkaline solutions (sodium hydroxide content greater than 1M) and the phosphate buffers but was not applicable for several of the solutions. The borate buffers became cloudy very early in the decomposition and with sodium hydroxide solutions less than 0.5M, precipitation of Fe (III) interfered. tior Wit} Fe ( corr by c spec of F tion from 26 In the more concentrated sodium hydroxide solutions, Fe (III) remained in solution and for moderate concentra- tions, did not contribute appreciably to the absorbance. With the phosphate buffers used in this study, the Fe (III) complex did contribute to the absorbance and corrections were necessary. Corrections could be made by detenmining the molar absorptivity of the Fe (III) species separately and calculating the contribution of Fe (III) at each point for the run. An approxima- tion of the ferrate concentration can be obtained from the following information. 1~ Atotal Are (VI) + AFe (III) I €Fe (VI)bCFe (VI) “’eFe (III)bCFe (III) 2- AFe (v1) = Atotal - 6s. (III)bCFe (111) = Atotal ‘ €Fe (III)b[CFe ( I) " CFe (VI)] initial concentration of Fe (VI) at t equals zero where C0 3. Substituting and rearranging, C31:"e(v1) =W " 6 bco Fe(VI) -.€re(III) b 4. Assume at t equals infinity, CFe(VI) equals zero, then 0 -— f Atotal " €Fe(III)bCFe(VI) ' 0 and 5° €Fe(III) :g—g-é—(VI) e where Ago equals the absorbance at the end of the decomposition. 27 6. Thus, Atotal " Aoo CF VI - e( ) €Fe(v1)b ‘ A00/‘33e(v1) 7. From (2) and (3) or from (5), a mere useful form is obtained. - A - Aoo AFe(VI) [W where A0 is the absorbance at t equals zero. 8. For each run, 1 - Ago/A0 is constant and Are (VI) 3 w "constant" The above values must be corrected for any absorption by the solution prior to the addition of ferrate. The assumptions made in developing the above equation are: (1) That absorbance other than solution blank is due only to Fe (III) and Fe (VI). (2) That the difference between the absorbance after decomposition and the blank absorbance is due to Fe (III). (3) That the absorbance first recorded, corrected for solution absorbance, is due to Fe (VI) alone. This is a good approximation for the kinetic runs in which the contribution by Fe (III) to the absorbance is negligible in the early part of the decomposition. Kinetic R The employs a the test passage 0: passing t: to low ox; Exte studies we more direc method wil 28 Kinetic Runs Using the Oxygen Analyze; The Precision Scientific Galvanic Oxygen Analyzer employs a silver-lead galvanic electrode separated from the test solution by a polyethylene membrane which allows passage of gases, but prevents ions in solution from passing to the electrode. This electrode is sensitive to low oxygen concentrations. Extensive exploitation of this method for rate studies was not undertaken since other methods were more direct and reliable. The limitations of this method will be discussed later. RESULTS AND DISCUSSION garters pH or 1 and the E decompc hydroxi will n: tion st T decompo hydroxi with so which t to 7 M’ to SEVe U wEre St' T1 using t] °XYgen r ferrite Chromite in Stir: °Xygen a 30 General The stability of ferrate solutions depends on the pH or hydroxide ion concentration, the ionic strength and the temperature of the reaction medium. Since hydroxide ions are produced during the decomposition of ferrate, buffered solutions or sodium hydroxide solutions, whose hydroxide ion concentration will not appreciably change, must be used for decomposi- tion studies. 7 To minimize the ionic strength change during decomposition, solutions from pH 7.7 to 2 M sodium hydroxide were adjusted to an ionic strength of two with sodium perchlorate. In a series of experiments in which the sodium hydroxide concentration was varied up to 7 M, an attempt was made to adjust the ionic strength to seven by the addition of sodium perchlorate. Unless otherwise stated, all ferrate decompositions were studied at 40°. The decomposition of ferrate (VI) was studied using the modified chromite method, the volumetric oxygen method and by following the absorbance of ferrate (VI). Decomposition rates Obtained from the chromite method were independent of moderate variations in stirring rate. The rates obtained from the volumetric oxygen analysis were dependent on the stirring rate. 31 Presaturation of solutions with oxygen did not effect the decomposition rates obtained by the chromite method or by the direct photometric method, but presaturation with oxygen or nitrogen was necessary for Obtaining reproducible data from the volumetric oxygen method. Decomposition rates were Obtained from stirred solutions, except for the direct photometric studies or other exceptions listed. Unless otherwise noted, all solutions were presaturated with oxygen. For pH 7.7 to 2 M sodium hydroxide solutions, the maximum concentration range of the ferrate solutions whose decompositions are to be studied by the modified chromite or the volumetric oxygen method is limited by the capacity of the buffers or the sodium hydroxide initially present to maintain a constant pH. Changes in pH from 0.1 to 0.3 unit were experienced from ferrate 3 M. The lower ferrate 3 concentrations up to 6 x 10- concentration limit was set at l x 10' M for the above studies in an attempt to maintain similar conditions with respect to the formation of an Fe (III) phase in the solution. Decomposition of 4 x 10'5 to 4 x 10"4 M ferrate solutions was studied by the modified chromite and the direct photometric methods. In the higher ionic strength solutions, the ferrate concentration limit was extended to 2.5 x 10'2 M. 32 Decompositions were followed by the modified chromite and oxygen evolution methods. Decomposition of 4 x 10-5 to 4 x 10'4 M ferrate solutions was again studied by the modified chromite and the direct photometric method. The extent of decomposition which occurred during ‘ the dissolution of the solid ferrate varied from about 50% in pH 7.7 buffered solution to only a few percent in l M or greater sodium hydroxide solutions. Rate Studies in theng 7.7 - 10.0 Range, 10"3 M Ferrate (VI) Solutions The decomposition of 1 x 10- 3to6x10‘3M potassium ferrate (VI) solutions in pH 7.7 to 8.9 borate buffers was second-order in ferrate concentra- tion. Decompositions followed by the modified chromite method gave second-order plots for 85 - 93% of the data collected for each run. Graphical representation of the decompositions are given in Figures 2 and 3, and plots are typical for this series of experiments. The rate of decomposition in the pH 7.7 - 8.9 range decreased rapidly as the pH of the solutions was increased. Values for the pseudo-second-order rate constants obtained by the modified chromite method varied from 425 to 24 l/mole-min. A solid Fe (III) phase formed very early in these borate solutions and two phases were normally 33 .oosooz nonsense essence: or» as oosoaaou an m~.m nos os.a an acosusaom consumed as AH>V assumes no soaps-omsoooo .m museum aw an ow nuances mH m e o\ _ _ _ _ s _ + o «as n.m o~.~ .e m.mm m.m mm.m .m .\ .ya has m.h mo.~ .~ . A . emu m.n oH.m .a . . Q 8. lrl -cua- nusosum. nllmmll, moan H> on o . H a .mxax omouo>e .usH o o . Arm . O 6 1.] u n e O O . . llm o w .9 D II. qm O 9 v \\ ifs. m N e o H II 4.m 34 .oosuo: oussouso essence: as» an oosoaaom as oa.m can mo.m em udoauoaom consumed as AH>V cumuuom mo soaps-oesouoo .m museum can can emu newness ems to as o _ _ _ _ T d an.» H.o ~¢.~ .e o . ma.m ~.m om.e .m a . s.am m.m mm.m .m . . o . o on m.m ma.m .H . w 1 a-.sssua-uaoe-a ma noaxxu>oom . t . . .mxox 09325 .ucH u c a s 1 . \ \ 5 as l e. o \x \ \ a \ e \. e 1 \ \ O \. e 1 \ \ \ Q.\ S \ . . \\ \ N H j H °| L’U presen precip prior of dec follov a grea occurr indice near 1 35 present during the decomposition. Addition of freshly precipitated hydrous ferric oxide to the solutions prior to ferrate addition did not effect the rate of decomposition. Comparison of data for the pH 7.7 to 8.9 range followed by the modified chromite method showed that a greater deviation from second-order ferrate dependence occurred in the pH 8.6 buffer (Figure 3) and probably indicated a change in the decomposition mechanism near the completion of the reaction. Studies in pH 9.1 borate buffer (Figure 3) showed that a change occurred in the rate's dependence on ferrate after approximately 65% reaction for 4.9 x 10'3 M initial ferrate concentration. Higher concentra- tions sometimes showed a change earlier in the decomposi- tion, depending on the extent of decomposition during mixing of the solid ferrate with the solution. A change from a second-order ferrate dependence to what appeared to be a first-order dependence occurred for runs of this concentration range. However, the linear portion of log C/Co plots gave different slopes for different initial ferrate concentrations. Decomposi- tion in 2 x 10"3 M ferrate solutions remained second- order for almost the entire run. 36 Addition of Fe (III), either as a ferric nitrate solution or as freshly prepared hydrous ferric oxide, had a marked effect on the decomposi- tion rate in pH 9.1 solution. Addition of small amounts of Fe (III) resulted in linear log C/Co plots for the major part of the decomposition. Increasing the amount of Fe (III) added, caused an increase in the rate of decomposition. In general, the rate did not increase linearly with increased amounts of Fe (III) added, a relatively smaller increase in the rate being obtained with larger amounts of Fe (III). Large amounts of the oxide caused very rapid decomposition. These effects are shown in Figures 4 and 5. Since the solubility of Fe (III) is very low (10'5 M), Fe (III) that is formed in the decomposition or which was added to the solution was essentially present as a solid phase on which a heterogeneous reaction is likely to occur. That the presence of Fe (III) in the solid phase was essential for affecting the rate of decomposition was demonstrated when the addition of glass beads, which provided a more extensive surface area, did not alter the decomposition rate. Potassium ferrate showed a similar pattern of decomposition in pH 9.4 carbonate buffer (Figure 6). 0. 01 37 FIGURE 4. Decomposition of Ferrate (VI) in' Buffered Solution pH 9.10 as Followed by the Modified Chromite Method. ._ 0 ‘ C G _ 9 a O \‘ r ' . L— ... O _ \ to _ \ . \ _ . o ‘1 2 _ \ Int. Fe(VI) Average Slope min-1 _ M x 103 pH x 103 I 1* 5.60 9.3 -9.70 2 4.20 9.2 -3.12 * 0.6 mg Fe (III) Initially Added J l I . l 0 ' 200 ‘ 400 Minutes 38 -.oosuos-oussouso endeavor or» an oesoHHOE mm coauamomEoomn I “HHHV om mascamvsou haaoauwcH os.m mo sosuoaom oououusm as AH>C ousuuon no cosusnoQEoooo .m sundae oo . oe nausea: pm 0 A A _ O HoO .0 mm.o- ~.m ~.m o.o e H mm.a- ~.m o.e N m . mo.m- ~.m o.e a N o m.oal N.m o.¢ mm H 7.56 mm .3 x z AHHHvom m 1: INCH x Omoam mmmum>¢ AH> mm.ucH OE 0 O o O 0 .v If 0 C m 0 .IT a o O u a ll . O . 1' UH) -+ 4 Te. C_C 0.11t 39 \ FIGURE 6. Decomposition of Ferrate (VI) in Buffered Solution pH 9.4 as Followed by the Modified Chromite Method. O 0 d- C 0 J. . 0 Q -t a 2 CO 6 a 0.11. “F H d- e 4. Ar- ui- 0 —l)- O O A. O 4* Int.Fe§VI) Average Slope x 103- M x 10 pH min‘1 0 3.9 9.4 -7.15 O 14.8 9.4 -9.50 0.01 ‘ l l 1 g I 0 100 Minutes 200 Ferrat order decomp indica 50 - 6 C/Co P of fer showed the be slopes ferrat Fe (II rate a decomp ClOSelj higher at 250 an ion. to be : Compar; Strenst 9.13 at in both 4O Ferrate solutions about 2 x 10‘3 M showed a second- order dependence on ferrate for a major part of the decomposition. Higher concentrations of ferrate indicated a second-order dependence for the first 50 - 60% of the decomposition, giving linear log C/Co plots for the last part of the decomposition. For the pH 10 carbonate buffer, decomposition of ferrate followed by the modified chromite method showed a first-order dependence on ferrate except at the beginning of the decomposition. However, the slopes from log C/‘Co plots varied with the initial ferrate concentrations. Addition of small amounts of Fe (III) to 2 x 10'3 M ferrate solution doubled the rate and gave a linear log C/Co plot for the entire decomposition. The increased rate compared more closely to rates obtained from solutions with a higher initial ferrate concentration. Magee (22) studied the decomposition of ferrate at 25° in solutions with a pH of 9.2 and 9.6 and with an ionic strength of 0.150. He found the decomposition to be second—order in ferrate for this pH range. For comparison, a borate solution was made with an ionic strength of 0.1 and a pH value of 9.23 at 25° and 9.13 at 40°. The decomposition at 250 was followed in both stirred and unstirred solutions by the 41 modified chromite method. Results are shown in Figure 7. The decomposition was found to be second-order in ferrate for about 80%.of the reaction. The stirred solution gave a value approximately 10% higher. The lower value agreed well with Magee's highest value for this pH, especially in view of the difference in the ionic strength of the solutions. Decomposition at 40° in the pH 9.13,ionic strength 0.1, solution showed a marked change from that at 25°. A first-order plot was obtained for the last 70% of each decomposition (Figure 8). However, the rates for different ferrate concentrations did not agree, the more concentrated ferrate solution decomposing faster. There is little question that the change in the order of ferrate was a result of the higher temperature and the studies in the pH 9 - 10 range show that Fe (III) plays a major role in the decomposition at 40°, but apparently is less significant at 25°. The decomposition in pH 7.7 to 10.0 solutions was also studied by following the evolution of oxygen from the solution. Results from this method correspond to those found by the modified chromite method except that the rates Obtained were lower. Decomposition rates in the pH 7.7 to 8.9 range varied from 116 to 13 l/mole‘l-min'l. Data collected 42 - . .oonuo=.muasouno amamacoz man an cascades mm..a.o aumaouum cacoH .~.m mm coauaaom aoummmsm ca omm um AH>V mumuuwm mo coauamoasoomo .a smoon can _ oma quacas cm a o a _ _, . _ _ o.m m.m . o.¢ conufium uoc N H.oa.u|. mam o.¢ nmuuaum H u." HIGHEIHImHoEsHI mm moaszH>vmm . s . .mxox ommuo>m .ucH \ a. 59 1.... G ‘x O L: O O 0 O c . IT \ \c a 1. \ \ \ \ \ . 1: \ \ s \ m o \ \ \ Ir \ \\ 9 \ H -u \ \\ °l oo C_C 0‘01 Ii [Ill 43 FIGURE 8. Decomposition of Ferrate (VI) at 40° in Buffered Solution pH 9.1, Ionic Strength 0.1 as Followed by the Modified Chromite Method. 1.0 f __ s O _.. 8 O Q .g __ Co \‘ O 0.1 ~- 4_ Slope -L Int.Fe§VI) Average x 102. M x 10 pH min- 1 2.30 9.2 ‘1-01 __2 5.40 9.35 -1.74 0 0.01 g ! ! 4 O 40 Minutes 80 by thi last 1 tion 1 repres oxyger deviat decom; is Che evolu1 44 by this method generally showed more scatter for the last 15 - 20% of the reaction. In Figure 9, decomposi- tion in pH 8.25 and 7.70 buffers is shown and is representative of decomposition followed by the oxygen evolution method in this pH range. The deviation from linearity near the end of the decomposition for the 5.9 x 10"3 M ferrate solution is characteristic of this method. The effect of the stirring rate on the oxygen evolution method was studied in pH 8.6 buffer. When the buffer was not presaturated with oxygen, a slower decomposition rate was obtained with unstirred (except initial mixing) or slowly stirred solutions, then with very rapidly stirred solutions. Presaturation of the solutions with oxygen or nitrogen and stirring at a slower rate gave values that agreed with the rapidly stirred solutions which were not presaturated. Increasing the stirring rate in solutions presaturated with oxygen or nitrogen did not significantly increase the decomposition rate obtained. Decomposition data for the pH 9.10, 9.35 and 10.0 buffer solutions obtained from the oxygen evolution method paralleled that found with the modified chromite method, although the decomposition 45 .nonumz aoaasao>m nomxxo.unp an cascades mm m~.m can on.» ma coauasom sommsm as AH>V «assume no coauamomEoumo .m smoon on em ow mGHDCHZ m w o 3 J a w % fix l u o 0.0m m.m N.m v o.vv ¢.w m.m m .u H mm m.h m.m N a _Wh mad m.h om.N H a H answelalmaoela mm moa x z . o a on .4 .dxwx 09392 AH>vmh.ucH ‘ . a no t e s o e 11 m . O O O a ~ s\ e 1.. IN. \\. \fl 00 \ \ C \ \O \ § U . 1' m \ \ \ C \5 \. \ LT \ Q \\ O \ g \ I: h N o .c .r m a O 5 LI m rates ferrat using compli less t this f pH 9.0 The pH using used 11 early Solutic hYdrox; had the The 10, AlthOug dilUte Possibl 301Utio Compara pho'Pha Compar a) and Carl 46 rates were again lower. The decomposition of ferrate (VI) in buffer solutions was first studied using the oxygen evolution method and to avoid the complications of the pH 9 - 10 range, ferrate solutions less than 3 x 10’3 M were studied in this range. For this ferrate concentration, the decomposition for pH 9.0 - 9.5 was mainly a second-order reaction. The pH 9 - 10 region was studied more extensively using the modified chromite method. The interchangeability of the buffering agents used in the buffer solutions was investigated in early studies using the oxygen evolution method. Solutions of borate, carbonate, phosphate and sodium hydroxide were made so a solution of one reagent had the same pH as a solution of another reagent. The ionic strength of the solutions was kept at two. Although some of the solutions were poor buffers, the change in pH was small for decomposition of dilute ferrate solutions (2 x 10‘3 M) and it was possible to compare the rates in the different solutions. Borate and carbonate solutions gave comparable rates at the same pH value. At pH 11.2, phosphate and sodium hydroxide solutions gave comparable rate values. However, at pH 10, phosphate and carbonate solutions gave different decomposition 47 rates. A faster rate of decomposition was found in the phosphate solution and was first-order in ferrate during most of the run. The difference in the decomposition rate of ferrate in phosphate and carbonate solutions is probably a result of the increased solubility of Fe (III), (10’3 M compared to 10'5 M for carbonate) in the phosphate solution. Since the phosphate solution is a poorer buffer at pH 10, the carbonate solution was used for the studies at this pH. The pH 9.13 (ionic strength 0.1) buffer was used for a single volumetric oxygen run. Results were similar to those obtained by the modified chromite method. The decomposition rate also agrees with the value obtained by the modified chromite method. Evolution of the experimental rate constants showed that the experimental second-order constants for pH 7.7 - 9.5 range were not the same for different ferrate concentrations in the same buffer solution. However, the data were reproducible for the same initial ferrate concentrations. Measurement of the pH value of the solutions after each decomposi- tion showed the final pH value varied a few tenths of a pH unit from the original pH, depending on the initial ferrate concentration. 48 Figure 10 shows that the variation of the experimental rate constants for the same buffer could be attributed to the slight variation of pH during each decomposition. The extent of pH variation was dependent on the initial ferrate concentration. An increase of two to three tenths of a pH unit reduced the rate 50%. ‘With the assumption that the rate is a function of the hydrogen ion concentration, specifically that, - 2- n Agiill). = K (Fe04) (H) . a plot of log kexp. versus pH was made, using the average pH value of the run. Figure 10 shows the linear relationship with a slope of -l, which corresponds to a value of 1 for n and indicates a first-order dependence on the hydrogen ion concentra- tion. Thus, the overall decomposition is third-order. Since a considerable amount of decomposition occurred during dissolution of the solid ferrate in the lower pH range, the actual pH change during the collection of data was smaller than the overall change in pH and linear plots were obtained for each run. The dependence of rate constants on pH, initial ferrate concentrations and third-order constants are given in Tables I and II for the modified chromite and oxygen evolution methods. Values of 1.97jt 700 100 ~ exp. 10 49 FIGURE 10. Variation of kexp. with pH of the Solution for pH 7.7 - 9.5. 700 —— Range Buffer Initialng 1 A Borate 7.7 B Borate 8.25 7’ A C Borate 8.7 O D Borate 9.1 E Carbonate 9.35 2 100 ~—- I exp.-- | l . I 10 .... 1 Modified Chromite d__ Method S : 1.12, n : 1 ~- 2 Oxygen Evolution Method l I l —— l I I 1 I I l I l I L 7.5 8.5 pH 9.5 50 10 0.30 x 10 and 1.36 t 0.18 x 1010 12 -mole"2-min-1 were obtained for the modified chromite and oxygen evolution methods respectively. The pH value used for calculating the third-order constants was important, a variation of 0.05 pH unit making a significant difference in the value obtained. Since a change of 0.1 to 0.3 pH unit occurred during the decompositions, it was necessary to use an average pH value in calculating the rate constant. It is likely that part of the variation in the values for the rate constant resulted from the uncertainty in the pH value used. The difference in rate values obtained by the modified chromite and oxygen evolution methods is apparent in Figure 10. For pH 7.8, the oxygen evolution method gave a lower value than expected. Other data also indicate that the oxygen evolution method is of limited use at higher decomposition rates. For the pH 9.0 - 10.0 range, the final pH values of the decomposed ferrate solutions also depended on the initial ferrate concentration. Part of the difference in slopes for the linear portion of log C/co plots may have resulted from the slight difference in the pH values of the solutions studied. 51 om.H o.mH mh.m H mH.mlmm.m 0N.N «m5.d o.NN «0.5H oa.m “0H.m N ma.mlmm.m memsm Quauom Oh.m mm omuuaum do: coHusHom MH.~ mm om.~ H ow.muom.m swans HHHHC as m~.m ~OH mm.~ H o¢.muom.m oH.~ m.Hn mm.¢ H mm.muo¢.m am.mm mm.m oo.~ NEH ummH m.oo u~.Ho u86 “mm.m m oo.muod.m oo.~ h0H.~ om «am oa.m ava.m m om.muom.m ¢~.~ “om.~ mm hNHH om.~ am~.~ N ov.muom.m ummmam muadom m~.m mm omens HHHHV ma om.m oHa mm.~ H om.anoa.a mm.H u3H mmm ammm mm.m “8.0 m mo.muom.a om.~ mNm ov.¢ H mm.auma.a mm.~ “am.~ mmd aaHe mm.~ amm.~ m om.anos.a udmmam madman on.n.mo mmmmmmm anHe-~-oHoeu~w. HutHEIquMoEuH MOszHH>Hmm mmmmqmm. omens ma oHuonx xmx condo: muHEouao oonHdoz was an dm3oHHom ma HH>V manuamm mo doHuHmoQEoomo .H «Haws I! mxuacam HucHEumuwHoEINH HIcHEuH-mHoEuH monZCEmE mczm .oz omcmm mm .. .me oHuonx H UQSCfiuCOU .H QNQNB 52 3.6 H 3H CflH no flow HmHHm HOW mm.H Om.® mHom H m¢.mlmm.m G5H MO 800 umuflm Rom Nm.H 05.0 fim.m H m¢.mlmm.m can mo ROM #mHHM Mom 0m.H Om.m OH.® H mm.mlmm.m and do mam umuHu nod OH.~ ma.m om.¢ mm.mumm.m H0MM5M wHMCOQHMU mm.m mm can no xmo umuHu uom Uwflfim mflmmfl mmmHU Vm.H mH.m moov O¢.mlmH.m can No *Oh vaHw HOW hm.H mN.® mH.¢ OfiomlmHom CSH m0 Xbm UmHHN Mom 0m.H flm.h N¢.N Om.mlONom Cflu M0 £60 umHHm Mom ¢¢.H mh.m 0m.¢ H O¢.mlmH.m fifiu NO ROB HMHHN “Oh mm.H mHom MN.N H Om.mlON.m Cfiu m0 ROB UmHHm Mom OV.H mm.m N¢.® H O¢.mlmH.m Hmwmflm mvmuom OH.m mm omens HHHHV mm mm.H o.m~ oH.~ mm.muom.m mb.H m.HN fim.N mm.mlom.m Hm.N h.Hm mm.m oo.mlom.m GM.N m.mN ww.m ocomlom.m a HucflEnNImHoEImH HIGHS: «WWW—Mia moaszHme 95% .oz mmcam mm OHuOHxs x dmddHudoo .H mHnue mxumeom monHucHE aux 2x2 Cash WEE dim Mama: 2% m nozcaucou .H ansE 53 coauummu mo mom pooH nos cowuoomu mo mom dooH dos couuomou mo mom dooH nos cowuomou mo som pooH nos ooooo HHHHV on us ¢.m oooom HHHHV on me a cocoa HHHHV on as am oooom HHHHV on us 0.0 poops HHHHV on ma H coauommu mo mom umoH Hos ooooo HHHHV on me o.o mxumEmm mH.H ho.m mo.H mH.N oo.m OH.® fim.m om.¢ mo.m OH.h m.mm ON.H mm.H m¢.H m¢.H MN.N Homummnoumoonmoo mm.m m0 oo.¢ oo.¢ oo.¢ om.m om.~ oo.m Nv.® om.m LII+1qxu|. NOHxHucHe H omscaucou .H magma nommom oumuom 0H.m mm monz HH>eom moss .oz m¢.m|mm.m oo.mlmm.m om.mlmm.m om.mlmm.m mN.mI0H.m mN.mIOH.m mN.mIOH.m MN.mIOH.m NN.mIOH.m om.mlmH.m mm.mImH.m mm.mImH.m omcmm mm oxomeom monanHeioxox moi: time. 35:62 doses as UUJCHSCOU .H moan-t... 54 coauHmomEoooU nouns wouuoooo soaumuamwuoum o: “.8338 ouodmooad mo.m mo.~ H wanna» oEmumn coauaaom HHudo cocoa .HHHV om ~o.~ mm.~ H omm. mm.m H v~.H ao.¢ H oowm mo.m H dommom ouoconuoo o.OH mmw mHHoEom NOHancHe.»mmum. monz HH>Vom odom.oz ooooHuooo .H oHHma mH.0HIo.OH HH.OHIO.0H OH.OHIO.OH mH.OHIo.OH ON.OHIO.OH mmcmm mm wawsm 0&6h0m ON A I IllIII I‘llllrlllilll mxumem chHetmuoHoEI H IIWIIIIIII [I’ll/If! . I HE: 10 OE: I . (Ii/Al! IIIHI/ oTonsm H H .deooH. monzHo> 83 ocoz oz moi... m0 COHUSHO>H 500%th 5n UG3OHHOE mm. fiH>v muUflhhavnu HO C0H.ufi.w.OQ.EOUva .HH .Ufinflflho 55 m¢.H o.mH «m.~ m¢.H m.mH ~¢.~ m mm.muom.m mm.H m.mH mH.m mm.H m.oH om.m m mH.mumm.m mmmmmmnmmmmmw o~.w mm am.H o.m¢ om.¢ H¢.H ¢.¢¢ mm.¢ m mm.mno¢.m mv.H o.¢o mm.~ mm.H o.mo mm.~ m o¢.muom.m summon ououom m~.m m ome.o oHH mH.m omm.o ooH sm.~ m om.anoa.a mmo.o o.~m ma.m OHo.o o.mm , mm.m m mm.auma.a ummmom common oa.o we I. - OszHo>|Hasv masm.oz omcmm mm mxumfiom HIGHEINIoHoELNH chafiuauoHoEIH OHIOHXM .. m mxmx 1|. I coauoHo>m comhxo afi wozoaaom mmAH>v oumuumm mo coauflmomEooon .HH magma 56 HowoSHoxo muommsn- om.m om.m mm.¢ om.m hH.N mo.N mGOHuSHom wwwwsm mo comHummEoo om.N om.N 0.0H HN.h mfl.o \I)‘l nommom ouodonuoo mm.m mm oo.N mm.N 0H.N a.a as some mooHo>e mH..H om.H can no xmb umuHm tom :OHusHom oumuom 0H.H can mo mom umuHm How.d0HusHom mvmcmmocm mo.H sou mo Rem aumuflu mom mH.H mo.H hh.H mm.H vH.H mxumaom H IcHEINIoHoEINH 0Huons IUH OEIH a . mxmx noodHucoo .HH oHnme nowmsm mummom oH.m Mm 3520?. 83 mcsm.oz om.mlo¢.m ow.mlom.m omomlmm.m om.mlmm.m om.mlom.m mmcmm mm 57 However, Fe (III) still must be considered an important factor in the decomposition of ferrate in this pH range. Attempts to relate Fe (III) and ferrate (VI) concentrations in a single rate expression for the ferrate concentrations studied between pH 9 - 10, were not successful. The usual practice of swamping the system with a variable to maintain an approximately constant value was not applicable here. First, large amounts of Fe (III) caused rapid decomposition and secondly, the reaction was heterogeneous and the surface area of the Fe (III) was important. For ferrate solutions to which Fe (III) was not added, the extent of decomposition during dissolution of the solid ferrate (VI) varied with the amount of solid ferrate present and provided a slightly different environment almost from the start of the decomposition. A final complication was the change in the order on ferrate which occurred during the decomposition. Some of the decompositions in pH 9.05 borate buffer solutions, not containing additional Fe (III), could be described fairly well by a 3/2 order dependency on ferrate (VI). However, this dependency does not hold for the pH 9.35 or 10.0 solutions, the pH 9.13 (ionic strength 0.1) solution nor solutions 58 that were treated with Fe (III). It is felt that the 3/2 order dependency was prdbably an "averaging" effect for some runs in this one solution. Decomposi- tion paths involving both first-order and second-order ferrate dependencies are prdbably operative in the pH 9 - 10 range, with the first-order path requiring the presence of Fe (III). This possibility is supported by the fact that a change occurs from second-order dependency on ferrate for the pH 7.7 - 8.9 range to a first-order dependency for solutions above a pH of 10.3. Also, values for the second- order decomposition rates obtained in pH 9.05 and 9.35 solutions gave values which agreed with those expected for this pH range. Rate Studies in pH 10.3 to 2 M Sodium Hydroxide Solutions, 10‘3 M Ferrate (vi) The decomposition of potassium ferrate (VI) was studied in phosphate buffers pH 10.3 to 11.1 and sodium hydroxide solutions from pH 11.9 to 1.9 M. The initial ferrate concentration was varied from 1 x 10"3 to 5 x 10‘3 M. The decomposition was first-order in ferrate and above pH 10.3 was not effected by the addition of hydrous ferric oxide or ferric nitrate. The same experimental rate constants were obtained when different initial ferrate concen by the order 1 graphs increa: increa: obtains 0.2 to decomp: is give range. T in pH 1 initial approxi ferrate Constan tiOn, t range, that th buffer 10‘3 M near tht C°ncent. 59 concentrations were used. Decompositions followed by the modified chromite method gave pseudo-first- order plots for at least three half-lives and many graphs remained linear for the entire decomposition. The rate of decomposition in the pH 10.3 - 11.1 range increased uniformly as the pH of the solutions was increased. Values for the experimental rate constants Obtained by the modified chromite method varied from 0.2 to 0.5 min-1. Graphical representation of decomposition in pH 10.3, 10.6 and 11.1 buffers is given in Figure 11 and is typical for this pH range. The experimental rate constant for decomposition in pH 10.3 buffer varied slightly with different initial ferrate concentrations. The variation was approximately t 10% for a three-fold variation in ferrate concentration. A lower experimental rate constant was obtained with a higher ferrate concentra- tion, the opposite of the trend found in the pH 9 - 10 range. Visual observation of the decompositions showed that the solubility of Fe (III) was much higher in this buffer than the buffers previously used. For 2 to 3 x 10'3 M initial ferrate solutions, precipitation occurred near the end of the decomposition, while in more concentrated ferrate solutions, precipitation occurred 60 6933 338:8 oonHooz ofi an 3333.... mm H.HH use b.0H .m.0H mm mummmsm sH AH>V oumuuom mo COHuHmomEooon .HH mmDon nouns 2 ¢H «H OH H o e N o J: + J H O H H J o O o I 3 5 O 2 a o .., 9 O 05H.o ¢.0H om.m m . Ho~.o b.0H o>.m N : mHm.o ~.HH mo.v H o . o HacHs mm mIOH x z /.., . . .933 ommum>€ :3 mm. .25 o/. olo Ea“... {ettal solid if 1" du: l.“ ‘ vaXUEE and 1 solu- solu' appr: dilut to ch with hydro of hy: neglig 12.8 ion. inc 61 during the run. The decomposition rate of a 2 x 10'3 M ferrate solution which contained a small amount of solid Fe (III), agreed more closely with the rate values for higher ferrate concentrations. Decomposition rates of ferrate (VI) in pH 10.6 and 11.1 buffers were unaffected by addition of hydrous ferric oxide and experimental rate constants were approximately the same for different initial ferrate concentrations. The pH 11.9 - 13.8 solutions are not buffer solutions, but sodium hydroxide-sodium perchlorate solutions. By studying more dilute ferrate solutions (2 x 10"3 M), the actual change in pH was limited to approximately one to two tenths pH unit for the dilute sodium hydroxide solutions and is comparable to changes found for the different buffer solutions with higher ferrate concentrations. As the sodium hydroxide concentration becomes greater, the contribution of hydroxide ion from ferrate decomposition becomes negligible. The decomposition rate of ferrate in pH 11.9, 12.6 and 13.00 solutions was first-order in ferrate ion. The rate decreased as the pH of the solution increased. Experimental rates constants as obtained by the modified chromite method range from 0.4 to 62 0.6 min-1. The effect of the ionic strength on the rate was investigated by adding sodium perchlorate to 0.1 M sodium hydroxide to give ionic strengths of 0.1, 2, 4 and 7. The solutions contained the same amount of sodium hydroxide, but had slightly different pH values ranging from 12.6 to 13.0. The decomposition rates of ferrate in the above solutions indicated a decrease in the rate with a higher ionic strength as shown in Figure 12. In pH 13.4, 13.6, 13.8 solutions and 1.30, 1.60, 1.90 M sodium hydroxide solutions, the rate of decomposi- tion decreased as the hydroxide ion concentration increased. Addition of a small amount of hydrous ferric oxide (10 mg) to the solution prior to the addition of ferrate did not effect the rate of decomposition. Typical plots for this pH range are shown in Figure 13. The decomposition of potassium ferrate was also studied by following the evolution of oxygen from ferrate solutions. Rate constants obtained from first-order plots were lower than those obtained by the modified chromite method. Data obtained by the oxygen evolution method gave kinetic plots which generally showed a short induction period and deviations from a linear plot were usually found after two or three half-lives had 63 .Uozuwz ouHEouno oonHooz onu an ooonHom mm HH>V muoouoa mo doom coHonoosoooo on» do Humdouum oHcoH on» mo Hoommm .~H mmoon mw SC 6 v u H: m o H .. i H i H H v no N H o o a. O I O 0 O o a. . 1 av . O a a mm.o o.mH oH.m o . a Ho.o a.~H ao.~ m m om.o o.~H ao.~ m . H.o ma.o o.~H ao.m H . numdouum HudHe mom, mLOH x z UHCOH ..mxox ommuw>< AH>V0h.ucH H.o 64 .oosuoz ouHsouno oonHooz ogu an ooonHom am odoHoaHom oonouomm eoHoom dH HH>V ououuom z muoH x a mo doHuHmomsoooo .MH mmoon oH mouocHz m o m a. 1 0a. 0 maH.o m.H v oo~.o o.H m ¢-.o m.H m . mm~.o Hm.mH one H Hun—HE z mowrxouomm o . .mxox EHHHoom O o O O I o O O O r O O T O _.l r1 ,l ..| .7. // T passe evoli and I with runs witho demon Pretr. potas: oxyger oxyger nitrOg With 0 nitrog OXYQEn rather by the had no 65 passed. Typical decomposition plots for the oxygen evolution method are shown in Figure 14 for pH 11.1 and 11.9 solutions. The effect of presaturation of the solution with oxygen can be seen in Figure 15 which shows two runs in pH 12.60 solution, one with and one without presaturation with oxygen. The plots demonstrated typical behavior of unsaturated solutions. Pretreatment of the solutions with small amounts of potassium ferrate to saturate the solution with oxygen was not as effective as presaturation with oxygen gas. Presaturation of the solution with nitrogen served the same purpose as presaturation with oxygen. The fact that presaturation with nitrogen had the same effect as presaturation with oxygen indicated that a physical change was involved rather than a chemical effect. This is also supported by the fact that presaturation with oxygen or nitrogen had no effect in the modified chromite method. Increasing the rate of stirring when following decompositions by the oxygen evolution method gave only a slightly higher experimental rate constant and the decomposition plots were similar. Rapidly stirred solutions gave a rate constant up to 10% higher than the rates obtained with the regular stirring rate. 66 FIGURE 14. Decomposition of 3 x 10'4 M Ferrate (VI) in pH 11.1 and 11.9 Solutions as Followed b n E olution M hod. 0.0J S =-0.100 mig' sp- -0.105 min-1 Average kexp., pH min‘1 1 11.2 0.242 2 11.9 0.230 j l ‘ l 1 Division = 10 minutes FIG' 1.0 r 0.1H 0,o\ 67 FIGURE 15. Decomposition of Ferrate (VI) in pH 12.6 Solution as Followed by the Oxygen Evolution Method - Effect of Presaturating the Solution with Oxygen. 1.0 1 0D 3‘. .5. H . ‘ Q o " H o Vho-‘I 9 Vbb"vb o O D . 0 0.1 H . ° _ _ o ’ "‘ O a 0 G O _ 2 O H \\ Int.Fe(VI) _ M x 104 l 3.0 - presaturated with 0 2.5 - no presaéuration H ' 1 O 0.01L .L e 4 Jr 0 20 40 Minutes 68 The difference in rates obtained by the modified chromite and oxygen methods was not the result of the stirring rates used for the oxygen evolution method. Decomposition plots obtained by the oxygen evolution method generally showed a deviation from linearity near the end of the decomposition. This deviation is probably a result of the solid Fe (III) phase present in the solutions. Data obtained from decomposition of higher initial ferrate concentrations showed a larger deviation as shown in Figure 16 for 1.30 M sodium hydroxide solutions. These effects were characteristic for decompositions followed by the oxygen evolution method in this pH range. In addition to the lower rates obtained by the oxygen evolution method, a comparison of the data obtained by the two methods show the following differences. Values for the experimental rate constants obtained by the modified chromite method increased for the pH 10.3 to 11.9 range and then decreased as the hydroxide ion concentration increased. Values of the experimental rate constants Obtained by oxygen evolution show an increase for pH 10.3 - 11.1 range, but remain approximately the same through pH 13.4 and then decreased with higher 69 .FIGURE 16. Decomposition of Ferrate (VI) in 1.3 M Sodium Hydroxide Solution as Followed by the Oxygen Evolution Method. 1.0 Vaa- V \ 1 ‘4 Int. Fe(VI) kexp” O Mx104 min-1 1 2.5 0.126 2 6.0 0.125 0.01 % £ % i 0 20 40 Minutes 7O hydroxide ion concentration. Apparently there was a maximum rate that could be obtained when employing the oxygen evolution method. The high salt concentration of the solution does not appear to be the only factor responsible for the slower rates evaluated by the oxygen evolution method. By adjusting the ionic strength of 0.1 NaOH solutions from 0.1 to 7.0 with sodium perchlorate, an indication of the change of rate with ionic strength was obtained. The decrease in rate with increased ionic strength was not as large as the difference between rates found by the two methods. Also, a similar decrease in rate with increasing ionic strength was found with the chromite method. A pH against log kexp. plot, using the experimental rate constants obtained by the modified chromite method, is shown in Figure 17. The values for pH 10.3, 10.6 and 11.1 solutions gave a staight line with a slope of 0.59, indicating a dependence on hydroxide to the one-half power. However, if the first-order experimental rate constant (k = 0.0562 min'l) from the single decomposition run in pH 10.0 phosphate solution were included in the plot, a curve would result and the best straight line drawn through the points approaches a slope of one. It is believed that 71 10 .oonuoz ouHaouso oonHo sodas momma «H s 0H as you doHuoHom on» no mm,nuH3 exox mo :oHuoHuo> .aH mmoon mm flH IMH NH HH OH 0 H H H . - HoO 1 \ \\ , \ Holt \ \ mm.o u m T H N. I a X a“ .m To 11 u _ Th \ I I. T T Iii O.H 72 the rapid increase in the decomposition rate_with increasing pH may not be linear and the dependence on hydroxide ion may not be constant for this pH range. Between pH 11.3 and 11.9, the decomposition rate appears to reach a maximum value as shown in Figure 17. From pH 11.9 to pH 13.8, the decomposition rate decreases with increasing pH of the solution. A tangent drawn to the curve indicates a slope of-- 0.32. For more concentrated sodium hydroxide solutions, a plot of log kexp. versus log (OH) gives a straight line with a slope of -0.44, suggesting an inhibition by hydroxide ion to the one-half power (Figure 18) Experimental rate constants Obtained from both methods are plotted and both give straight lines with the same slope, although values obtained by the oxygen evolution method show considerable scatter. Data for the decomposition of ferrate in pH 10.3 to 2 M sodium hydroxide solutions are listed in Tables III and IV. 73 . .wuHHmHoz oonoupam EsHoom mo GOHuocsm mm .mxox mo QOHHMHum> .mH mMDUHm mmsHm> vogue: cOHusHo>m comuxo N mmsHm> tonne: ouHEou£O.pmHMHpoz H L l I 74 Hoe OHV cocoa HHHHV om oHHom summons oHHom HHHHV we no unsoEm HHmEm mmeEmm mm th.o NmHm.o om.¢ hmo.¢ NNm.o ON.N nommom oomsomona H.HH mo woo. H Homo mmN.o uHoN.o No.v uOh.m mmN.o NN.m moN.o hmoN.o om.N uNm.N uommom oumnmmond 6.0H mm mHo. +.¢hH.o th.o om.m omH.o u00.70 o¢.¢ H0H.m h©H.O mm.N mmH.o ummH.o OH.N hmo.N dommom ooonmmono H.0H mo ICHEH.QX0¥. m:_u~ z ._~:um H oosuoz ooHEouno ooHHHooz one an ooaoHHoa HH>V oumuuom mo COHuHmomEoumn .HHH mHnme N m.HHIH.HH H N.HHIH.HH N m.OHI©.OH H m.OHI®.OH N h.0HI®.OH H ¢.OHIm.OH N m.0HIm.OH H v.0HIm.OH N ¢.OHIm.OH adj odd: 75 moo. H ~36 Hm: oHv oooom HHHHV om oHHom oHo.o mm.~ mHo.o Hm.~ uoHoo umovo “oo.~ h$8 coHuoHom oononosm eoHoom o.mH m0 on.o omm.o hoom.o oa.H aSH coHosHom oonooosm aoHoom o.~H mm ooo. HHHoo oHo.o mo.H uo~o.o n~oo.o a-.H N3H coHooHom oonooon eoHoom o.HH m0 , ooo. HmHmo Hoe oHV oooom HHHHV om oHHom mHm.o om.m mmeEmm |CHE..mex mOHx E HH>Mwm H ooocHocoo .HHH oHnme mcsm Doz H.mHIo.MH H.MHIO.MH N.NHI®.NH o.NHIm.HH m.HHIH.HH mmcmm mm 76 moo. H ammo hNN.o MN.N NONN.o uVNN.O xHm.m Hoh.m m COHpsHom mUonupmm EDHUom 2 0m.H moo. H mmmo mmN.o oo.N u:uN.o hmmN.o hSwim NOh.m m um.MH doHuoHom oonouosm eoHoom m.mH mm ooo. H Homo Hoe oHV pmppm HHHHV mm UHHom mNN.o mo.m H no.mH th.o uth.o mv.m u23m N no.mH COHusHom mUonuomm EsHoom o.mH mm moo. H ommo mNm.o wN.N ummm.o u0mm.o uN©.m xmh.m m Iv.MH coHusHow oponupmm ESHUom «.mH mm mmeEom IGHE..mxmx mOHX ZIHH>Vom mcsm .oz omcmm mm H ooodHucoo .HHH oHnoe 1.“ . 77 o.a nomcoaum oHcoH o.e Hoodooum oHcoH H.o Hoodooum oHooH mNm.o mNm.o uNNm.o mH¢.o ¢N¢.o uNH¢.O om.o Nm.o nmh.o mo.N HH.N OH.N “ao.m sumcmuuw UHCOH mcHNHm> I mCOHusHom mponupmm EsHpom Hoe oHo ooooo HHHHV om oHHom mxumEmm moo. H maH.o MbH.o u00.70 hhhH.o um¢.N Nm.H "mm.m coHusHom monouowm EsHoom 2 om.H ooo° H.mo~.o OHN.o ®ON.O uOON.O HucHE .mxmx no.0 0N.N hNm.m ooooHodoo .HHH oHnoe OH X 2 ~H>H0m IO.MH Ih.NH Im.NH mczm .OZ umdmm,@m 78 mm~.0 uOMN.O h®.N uHo.N N doHooHom oonouosm eoHoom o.HH m0 ooo. Hommo OMN.O xmmN.o m¢.m kom.m N N¢N.o mm.N H COHuflHom mumnmmonm H.HH mm moo. HomHo NMH.o Hwom NhNH.o umNH.o xHm..N hoooN m ooHuoHom oumnooood o.oH mm mHH.o NNH.o m©.N H oHH.o o¢.m H ooHuoHom oposmmond m.oH mm mxumEom cHE..mxmx MOHMZ HHMth mcsm .oz vogue: COHusHo>m-comhxo onu Hal an UOBOHHom mm AH>V oumuuom mo COHpHmomfioomQ .>H OHAMB o.NHIm.HH m.HHIH.HH N.HHIH.HH h.0HI®.OH ¢.OHIm.OH m.OHnm.oH wUGm m ma 79 mMN.o om~.o uommo Hm.m “mo.~ :oHuoHom oonooosm eoHoom «.mH mm mmm.o mmN.o uomm.o mm.N hN©.N coHosHom oonoaosm soHoom o.mH mm o.o Homemaum oHdoH on.o o~.~ oH.o ouooooum oHooH Hom.o a¢.m umcmmum OHcom mchHm> n msoHusHom mpHxOnowm EoHoom ooo. H mmm.o N2 ouHs noumuoummwum cOHuoHom qNN.o o¢.m ummN.o umemo uH.o.N kmo.N ooHooHom oonooosm eaHoom o.NH mm ooo._H Hmm.o mHN.o o¢.m mHomEom HucHe..mxox monz HHsHom ooocHuaoo .>H oHnoe N I¢.MH N IO.MH H m.NHIm.NH h.NHIw.NH r-il H N.NH mcnm .oz wmsmm m0 80 OHH.o mOH.o uNHH.o m¢.m uom.N COHuOHom OUHHOHUNm EsHUom z om.H NMH.o omH.o ummH.o om.m um©.N cOHuoHow oonoHowm EsHoom z oo.H oNHoo hNH.o u©NH.O ¢¢.m um©.N nOHusHom OUHXOH©>E ESHUom 2 om.H oaH.o oaH.o hEH.o mo.m hom.m ooHoaHom oonooomm eoHoom o.mH mm moH.o mxumEmm moH.o uNo.70 oo.N hmo.N ooHuoHom oonoaon eoHoom o.mH mm CHEi.meM mOHXE AH>vwm HI ooooHocoo .>H oHnoe (\i Im.MH I©.MH MCDNH .OZ mmcmm mm 81 cOHusHom oumnmmonm mponupmm Echom COHusHow oumconumo o.oH mm COHuoHom mumzmmosm mmeEom hmH.o HO.N NmH.o MH.N moo.o NMHo.o no.N h0min omo.o mm.H mCOHusHom Hmmmsm mo comHummEou aHe..mxox onz HH>oom H- . m ooodHucoo .>H oHnme mcom ooz N.HHIO.HH H.HHIO.HH H.0HIO.OH H.oHao.oH 002mm mm 82 Decomposition of Dilute Ferrate Solutions, pH 7.7, 2 M Sodium Hydroxide Solutions As a result of the higher ferrate concentrations used for the decomposition studies described in the previous section, an Fe (III) solid phase formed very early in the decomposition. The study of the 5 to decomposition of dilute ferrate solutions (3 x 10- 1 x 10'3 M) had the advantage that the decomposition could usually be followed in a homogeneous system. For many of the solutions, the ferrate concentration 5 M to 2 x 10’4 M if had to be limited to 3 x 10‘ the solution was to remain homogeneous. Only an approximate decomposition rate could be obtained for solutions in which ferrate decomposed rapidly or in which precipitation of Fe (III) still occurred. The pH 10.0 - 10.6 phosphate buffers and the 1 - 2 M sodium hydroxide solutions had the best properties for studying the decomposition in a homogeneous solution. An estimate of the solubility of Fe (III) in the solutions used in the decomposition studies was made by allowing a ferrate (VI) solution to decompose at 40°, filtering the warm solution through a fine-sintered glass filtering funnel and determining the iron content of the solution spectrophotometrically using 1, 10 - phenanthroline (35) as the reagent. With 83 the exception of the phosphate buffers, the analysis of the filtered solutions indicated that the solubility of Fe (III) was less than 1 x 10"4 M. In pH 10 - 10.6 phosphate solutions, the solubility of Fe (III) was 5 x 10'4 to 1 x 10"3 M, decreasing with increasing pH. Solubility in the pH 11.1 phosphate solution was lower than in the other phosphate solutions. Decomposition of dilute ferrate solutions was followed by the direct photometric and the modified chromite methods, except in the borate buffers and pH 11.5 - 13 solutions, in which the direct photometric method was not applicable. When the direct photometric method was used to follow the decomposition of ferrate, the molarity of the ferrate solution could be conveniently calculated from the absorbance of the ferrate solution. To evaluate the molar absorptivity of ferrate (VI), an aliquot from a ferrate solution was withdrawn from a spectrophotometer cell while the absorbance was being continuously recorded and the ferrate content of the aliquot determined by the modified chromite method. A value of 1100 $.40 l-mole-lncm“l was obtained with several ferrate solutiOns and was used to calculate the ferrate molarities in the kinetic runs. This absorptivity value agreed well with that reported by Wood (12). 84 When the decomposition of ferrate was followed in pH 7.7 to 9.1 borate buffers, the solutions became turbid in the early stages of decomposition. In an attempt to avoid the above complication, the decomposition was followed in carbonate solutions of pH 8.2 to 9.4 and in a phosphate solution with a pH of 9.5. In the pH 7.7 a 9.5 range, the decomposition of ferrate solutions less than 4 x 10.4 M appears to be first-order in ferrate. This is in contrast to the secondmorder dependence found for the more concentrated solutions described in the previous section. The decomposition of dilute ferrate concentrations in pH 9.5 to 2 M sodium hydroxide solutions also appears to be firstmorder, as found for the more concentrated ferrate solutions. Curve A of Figure 19 shows the decomposition of ferrate in pH 7.7 borate buffer. The change in the slope is characteristic for dilute solutions in the borate buffers and the change coincides with the appearance of a turbid solution. Decomposition of ferrate in the borate buffers usually give similar curves for 1 x 10-4 to 3 x 10‘4 M ferrate solutions in the same pH. The parallel curves for different ferrate concentrations suggest a firstnorder dependence on ferrate. 85 FIGURE 19. Decomposition of Ferrate (VI) in pH 7.70 and 8.25 Borate Buffers as Followed by the Modified Chromite Method. 1.0 i.- s S 0 : .0 e a i. \‘Q 1 ‘ ~ - v . . 1 0' ‘ O 2‘ C . . . . 1 . . 3 O H o O ‘ C § /c. ~ . 0 2 0.11 .J : 1 _ Fe(VI) Slope min-1 EH Mx104 Initial Final 1 1 7.7 1.83 -0.0850 -0.0550 2 8.25 2.38 -0.0780 -0.0255 _1 3* 8.25 1.95 --- -.0260 * Fe(III) added to run. 0001 : A 5. 1 0 15 20 Minutes 86 The decomposition of ferrate in pH 8.25 borate buffer is shown in Figure 19, curves B and C. Curve C shows the decomposition of ferrate in a solution that contained a small amount of Fe (III) solid prior to the addition of ferrate. Comparison of curves B and C shows that the change of slope found in B was due to the formation of an insoluble Fe (III) phase in the solution. For the pH 7.7 - 9.1 borate buffers, the initial slope of each curve was used to estimate the experimental rate constant for a homogeneous solution. The decomposition of ferrate in the carbonate and the pH 9.5 phosphate solutions was followed by the direct photometric method and dilute solutions of ferrate remained homogeneous throughout the run. Linear log A/Ao plots were obtained for two or three half-lives as shown for the carbonate solutions in Figure 20. Decomposition rates in pH 9.5 phosphate solution were comparable with rates obtained in the pH 9.4 carbonate solution, indicating the inter- changeability of the buffer solutions for dilute ferrate solutions. Decomposition of ferrate in the pH 9.4 carbonate solution was also followed by the modified chromite method and for the similar initial ferrate concentration, comparable decomposition rates were obtained by both methods. 87 FIGURE 20. Decomposition of Ferrate (VI) in pH 8.25, 8.65 and 8.40 Carbonate Solutions as 1 0 Followed by the Direct Photometric Method. . (a. . |\‘ s‘ . r.“ - a H N . H a ‘ H n O ‘ O H 0 . c O 0 O H O A/ . . . 9 A0 1‘ . H O 3 . O 0 O 2 0.1_ , a q 0 _ . Fe (Vi) k _ 1 pH MxlO exp.,min ‘ 1 8.25 0.96 0.140 - . 2 8.65 1.04 0.0540 3 9.40 1.16 0.0182 H O l 0 . 0 l e 4 4 .L 0 40 80 Minutes filfifitfilfi mini pug , H a... a. . 88 The study of the decomposition of ferrate in pH 7.7 - 9.5 solutions permitted only small variations in the initial ferrate concentrations. Even with the limited variation in the initial ferrate concentra_ tion, it was noted that the experimental rate constant varied with the initial ferrate concentration used, a higher rate being obtained for decompositions starting with a larger initial ferrate concentration. The difference in rate of decomposition with the initial ferrate concentration is discussed later in this section. In view of the experimental difficulties and variation in the experimental rate constants with the initial ferrate concentration, comparison of decomposition rates should be made only for rates obtained for similar ferrate concentrations. The decomposition rate of ferrate decreases with increasing pH until approximately pH 9.5, after which the rate increases. A comparison of the decomposition rate with the pH of the solution for pH 7.7 - 9.5 was made by plotting the log of the experimental rate constant against the pH of the solution. Rates from similar initial ferrate con- centrations were used for each comparison. Figure 21 shows that a linear relationship exists for decomposi- tions followed by the direct photometric method and for 1.0 .01 89 FIGURE 21. Variation of kexp,,for pH 7.7 - 9.5 range for Dilute Ferrate‘(VI) Solutions. « (D Absorbance -lx 10'4M Fe(VI) - Carbonate ~ Solutions (pH 9.5 value,Phosphate Solution) (D Modified Chromite Method - 2 x 10'4 M - Fe(VI) using Initial Slope - Borate Solutions 90 rates obtained from the initial slopes of plots of decompositions in borate buffers. Slopes of the log kexp. plot indicate a first-order dependence on hydrogen ion, as found in the previous section for 10—3 M ferrate solutions in the pH 7.7 - 9.5 range. Experimental rate values, ferrate concentra- tions and method of analysis are given in Table V for solutions used in the pH 7.7 - 9.5 range. The decomposition of ferrate in the pH 10 - ll range was studied in phosphate and carbonate buffers. The solubility of Fe (III) is much higher in the phosphate buffers and allowed a more extensive study of the decomposition of ferrate than was possible in the pH 7.7 - 9.5 solutions. The phosphate buffers were used to determine the effect of a soluble Fe (III) species on the decomposition rate of ferrate (VI). The decomposition rate of dilute ferrate solutions increases with increasing pH from approximately pH 9.5 to 12, as was found in the previous section with 10'3 M ferrate solutions. However, when the log A was plotted against time for decomposition of dilute ferrate in pH 10 to 2 M sodium hydroxide solutions, a linear relationship was obtained only after an appreciable amount of decomposition had occurred. Also the experimental rate constants obtained from the linear 91 .mxwx now owns omoam HMHuHcH .mxox now owns vQOHm HMHuHcH cowuaaom wanna» anon ou befivm AHHHV 0m mo unseen HHmEm .mxox Mom poms omon HmHuHcH .mxox you coma .omon HmHuHcH mxumEmm m «nHo. op noHo.o H eHmo. m ommo. ou mmHo. N mmo. ou mmo. H omo. on mac. m «mo. on «mo. H omo. 0» moo.o m oeH. 0» omH.o H memo. on ommo.o H oma.o ou hoa.o oHuumsouogm Ame muHsouno emHMHeoz .Hv * 0H.H 0» mm.o ¢®.H mm.~ ou @H.H H¢.H ou mh.o OH.m 0» v®.H mH.H ou ¢O.H m.¢ ou MH.N mm. 0» 0mm.o mm.a ou mm.H mmom 0» mm.m a 0mm. 0» mma.o mo.m 0» m¢.H t. mHmWHwCG HIGHEioQXGVH fiOHXSNHKC man no vogue: - m.m I 5.5 mm .mCOHusHom HH>V mumuumm musHHo mo coauwmomEoowa .> magma N H w mcsm.oz om.m mm.m mm.m omom oa.m m©.w meow mN.m m~.m mN.m 05.5 .mm oumnmmonm uumvconumo oumconumo mumconumo oumuom mumconumo oumuom mumsonumo mumuom mvmuom mumnom cOHusHom W" ._‘- .3 92 portion of the log A plots, varied with the initial ferrate concentration used, indicating that the decomposition was not a true first-order decomposition. Graphs of the decomposition in these solutions are presented as log A rather than log A/Ao to allow the better comparison of a series of decompositions in the same buffer. The decomposition of ferrate in pH 10.0, 10.3 and 10.6 phosphate buffers was followed by the direct photometric and the modified chromite methods. Figures 22 and 23 show the decomposition of ferrate in pH 10.0 and 10.6 buffers and demonstrate variation of the slopes with different initial ferrate concentrations. Decomposition in the pH 10.3 buffer was similar to the pH 10.6 buffer. The induction period is more pronounced for lower initial ferrate concentrations and as shown for the 4 x 10"5 M ferrate solution in Figure 23, it extends for over 50% of the decomposition. In general, the decomposition rate in these phosphate buffers increases with a higher initial ferrate concentration up to about 3 x 10'4 M after which the rate is relatively constant up to l x 10.3 M ferrate. As the initial ferrate concentration is increased beyond 1 x 10‘3 M, the rate begins to decrease until precipitation of Fe (III) occurs. 93 Minutes FIGURE 22. Decomposition of Ferrate (VI) in pH 10.0 Phosphate Buffer as Followed by the Direct Photometric Method. 1.0 - ¢ 1 -2\\ e s § .. 3 g - s..- ¢ s - 3 s s a e . . Q ¢ Q 6 § ‘ A“ e‘ . 3 ‘9 . a ' e = o o . 6 3 I e ‘ s ' e ‘ O. 0.1- a. ’ ‘. e ’ ,3 o 9 H :3 O . es - .. 9 - Int.Fe VI) A9 M x 10 Slope min-1 ” l .835 2.82 -0.0219 2 .578 1.95 -0.0198 - 3 .312 1.05 -0.0l6l 4 .178 0.60 -0.0ll4 § J l l O 20 ‘ 4d 94 FIGURE 23. Decomposition of Ferrate (VI) in pH 10.6 Phosphate Buffer as Followed by the Direct Photometric Method. Int. Fe(VI) ._§2___ MX104 Slope min-1 1 0.885 2.98 -0.212 2 0.610 2.06 -0.206 3 0.367 1.24 -0.156 .4 0.127 0.43 -0.103 n G a 0 0 9 o dr— 1 1 r Minutes 95 Using the pH 10.6 phosphate solution as an example, the experimental rate values from log A 1 plots increase from 0.2 min- at 4 x 10'5 M ferrate 1 to 0.45 min- at 2 x 10’4 M ferrate. From 4 x 10‘4 to l x 10"3 M, the value was approximately 0.36 min-1. Initial ferrate concentrations of 2 x 10‘3 to 5 x 10'3 M described in the previous section, gave relatively constant rate value of 0.26 min-1, but precipitation of Fe (III) occurred during the decompo- sition in these solutions. The decrease in the experimental rate constant with increasing ferrate concentrations found with the pH 10.3 buffer in the previous section, showed a similar trend as found in these dilute ferrate solutions and was probably due to the variation in the Fe (III) concentration in solution prior to precipitation. The solubility of Fe (III) in the pH 10.6 phosphate solution is lower and this trend is not as apparent. The effect of Fe (III) on the decomposition rate was studied by adding solid potassium ferrate to a decomposed ferrate solution. By following the decomposition by the direct photometric method, several decompositions could be run in the same solution. The concentration of Fe (III) present in the solutions was assumed to be equal to the total amount of _ . '1'7 1"" 96 ferrate (VI) that had decomposed in the solution. In addition to being convenient, this procedure maintained identical experimental conditions, except for the concentration of Fe (III) in solution. Analyzing the decomposed solutions from several direct photometric runs with the modified chromite method, showed that no oxidizing power remained in the decomposed solution, indicating Fe (III) was the only iron product. The importance of Fe (III) in the decomposition of ferrate is seen in Figure 24, where the decomposi- tion rate of successive additions of potassium ferrate was followed in pH 10.6 phosphate buffer. A much faster decomposition rate was found for ferrate solutions which contained Fe (III) prior to the addition of the ferrate to the solution. For the same initial ferrate concentration, a faster decomposition rate was obtained from solutions which initially contained a higher Fe (III) concentra_ tion. Further addition of Fe (III) would probably result in a decrease in the decomposition rate until precipitation of Fe (III) occurred, after which the rate would remain constant. Decompositions in solutions initially containing Fe (III) usually did not show an induction period, and if sufficient 97 FIGURE 24. Effect of Fe(III) on the Decomposition Rate of Ferrate (VI) in pH 10.6 Phosphate Buffer as Followed by the Direct Photo- metric Method. 1.0 _, Int. MFe(III) x104 Ao Int.FeiVI) Slope 1 or' -- M x 10 min‘1 a l ---- .365 1.23 -0.156 _ 2 1.23 .235 0.79 -0.187 3. 2.02 .405 1.36 -0.204 S . ‘ 9 3 ‘ a ' a l‘ 3 '3 ° 1 v o A .~ - ' Q 2 v A \‘ u - 0.1_ _ 0 2 9' '1 a - g C _ v . , .1 N “H ‘1' g E , \V .1 0‘ - \‘\ §\ — g‘. o -\ . O O (3.01 e = i 5 ° % 0 4 8 12 16 20 minutes 98 Fe (III) was present, linear plots were obtained for the entire run. To compare the rates of decomposition in the phOSphate and carbonate buffers, a series of runs were made in pH 10.0 carbonate buffer. Experimental rate constants obtained from the carbonate buffer agreed well with those obtained in the pH 10 phosphate buffer for similar initial ferrate concentrations. When decomposition rates in phosphate, borate and carbonate solutions at the same pH were compared in the previous section, a much faster rate of decomposition was found for the pH 10 phosphate buffer compared to the pH 10 carbonate buffer, although rates in borate and carbonate buffers agreed. The difference in the decomposition rates in carbonate and phosphate buffers is due to the difference in the concentration of Fe (III) in solution. With the 10'3 M ferrate concentration used in the previous section, the concentration of Fe (III) in solution is much higher in the phosphate buffer. With the dilute solutions studied in this section, all the Fe (III) formed during the decomposition remains in solution and the same decomposition rate was obtained in each buffer for a similar initial ferrate concentration. 99 There is no question that Fe (III) is involved in the decomposition of dilute ferrate solutions in the pH 10.0 to 10.6 phosphate buffers. It is very likely that Fe (III) is important in the decomposition of ferrate in all of the solutions from pH 7.7 - 2 M sodium hydroxide. Because of the low solubility of Fe (III) in the pH 7.7 - 9.5 solutions, the effect of a soluble form of Fe (III) in the ferrate decomposition could not be investigated, but it is possible that the variation in the decomposi- tion rate with different initial ferrate concentrations is due to the Fe (III) formed in the decomposition. For the 10‘3 M ferrate solutions studied in the previous section, precipitation of Fe (III) occurred early in the decomposition and maintained a relatively constant amount of Fe (III) in solution. For these solutions, the decomposition rate did not vary with different ferrate concentrations and linear plots were obtained when the log of the ferrate concentration was plotted against time. The amount of Fe (III) formed in the decomposition of ferrate is equal to the amount of ferrate (VI) that has decomposed, assuming that there are no stable intermediates. The fact that the same decomposition rate was Obtained by the direct photometric and the fismflglq HI. H kiwi s 100 modified chromite methods suggests that any inter- mediates formed are short-lived and their concentra- tions are very low. If the Fe (III) formed during the decomposition of ferrate remains in solution, it should be possible to relate the decomposition rate of ferrate with the concentrations of ferrate (VI) and Fe (III) present in the solution. Two experimental rate constants were found which, within certain limits, could be used to describe the decomposi- tion. The first equation assumed a first-order dependence on ferrate (VI) and Fe (III) or that, mfgewfil = k'CFewIfl [Fe(III)] . The second equation assumed a first-order dependence on ferrate (VI) and a one-half—order dependence on Fe (III) or that, -dl—§E(VIJI - k" [Fe(VIfl [Fe(III)]? Substitution of C for [Fe (VIE and (Co - C) for [Fe (III{] in the equations and integrating (37) gave 1n [co - c] C 3: tanh l[-oc J9] = (C0) k" t-+ constant respectively. Plotting log [Go— C' or tanh 1 Co k' te+ constant and _ 35 H against time should give a straight line if Co 101 the data fit the rate equation. For the first equation, an experimental rate, koexp., constant was obtained by dividing the slope of the line by Co 2.303. For the second equation, the slope was divided by.%? 2 to obtain the experimental rate constant, k"exp. It was found that for the pH 10.3 and 10.6 buffers, data up to an initial ferrate concentration 4 of approximately 1 x 10- M fit the equation which ‘assumed a first-order dependence on Fe (III). The 4 to 3 x 10'4 decomposition of l x 10- M ferrate solutions could be described by assuming a one—half— order dependence on Fe (III). For 3 x 10’4 to 1 x 10"3 M ferrate solutions, the decomposition rate appears to be first-order in ferrate and relatively constant slopes are obtained when the log C was plotted against time. For the pH 10 phOSphate buffer, only the data from an initial ferrate concentration less than 8 x 10"5 M appeared to fit the equations developed above and for these decompositions, the first equation was used. For decompositions in pH 10.3 and 10.6 phosphate buffers which initially contained Fe (III), the equation which assumed a one-half-order dependence on Fe (III) could be used to describe the system if the additional Fe (III) was accounted for in the 102 equation. This was done by using a C00 value which was the sum of the ferrate and Fe (III) concentrations initially present. However, the equation could only be used for limited amounts of Fe (III) initially present, after which a lower value was obtained for the experimental rate constant and a linear relation- ship was no longer Obtained. This is in accord with the previous Observation that the decomposition rate of ferrate did not continue to increase as the Fe (III) content increased. Figures 25 and 26 show the decomposition of ferrate in pH 10.3 phosphate buffers using the two equations which were developed. The plots are typical for the decompositions which can be described by each equation. For ferrate concentrations above 4 x 10"4 M or for solutions containing larger amounts of Fe (III), the decomposi- tions were plotted as being first-order in ferrate. Decomposition of dilute ferrate solutions in the pH 11 to 13 range is difficult to study because of the rapid decomposition rate found in these solutions. Precipitation of Fe (III) occurred in many of the solutions and a homogeneous solution persisted only for part of the run. The decomposition of ferrate in these solutions was followed mainly by the modified chromite method, but the direct photo- metric method was used for few solutions. FIGURE 25. 103 Decomposition of Ferrate (VI) in pH 10.3 Phosphate Buffer as Followed by the Direct Photometric Method - k.exp plot. 10 ' 2 I; '1 1 . .1 3 . ‘0 '1 O H O 9.2—‘3’ . . l.0~ - - o ' - k'exp., n ‘ M Fe(VI)x10 1-mole‘1..m;i_n'1x103 ‘ u l 4.34 3.01 ~_ 2 7.34 2.84 r '1 a O ’ O «1 .i' i (D e 0.1 O J. Jr 40 Minutes 104 FIGURE 26. Decomposition of Ferrate (VI) in pH 10.3 Phosphate Buffer as Followed by the Direct Photometric Method - knexp. plot. 2.8 3 2 2.41. 2.0-- 1.6-- tanh-l (. )5 ” A - A _ o _ ' 4 k“exp., M Fe(VI) x 10 115-“.018-‘5- min"1 0-3 4* 1.58 22.9 2.17 22.6 3.98 21.0 0 i + i i o 4 8 12 16 20 Minutes 105 A variation in the decomposition rate with the initial ferrate concentrations was again noticed for these solutions. Several decompositions showed an induction period as noted for the pH 10.3 and 10.6 solutions, indicating that Fe (III) was still playing a role in the decomposition. Due to the low solubility of Fe (III) and the rapid decomposition rate, only approximate values for the decomposition rate could be obtained. Values were obtained by plotting the log C against time and determining the slopes of the linear portions of the plots. No consideration is given to the Fe (III) formed and the values are given as first-order rate constants only for comparison purposes. Comparison should be made between decomposition rates for similar ferrate concentrations. Experimental values for decompositions of ferrate in the pH 10 to 13 range are listed in Table VI. For several of the decompositions in the pH 10.3 to 10.6 solutions, two experimental rate constants are listed, one calculated from the linear portion of the log A plot and the other obtained by using one of the two equations mentioned above. Values from log A plots for these decompositions are listed merely for comparison purposes. It is of interest to note that when a comparison of decomposition rates for pH 9.5 to 13 are made by 106 oumuamaooum no: 6H6 HHHHVom H memo.o ~.o~ H momo.o mm.H accumum HHHuHuHcH AHHHme ‘8 euon em.~ m omeo.o oe.H ucmmmum mHHmHuHcH HHHHme z euoH mmeo. mo.H x mo.H “mo.H m “3.3. hmmH omeo. NH.m Noomo. umo...” m “83.0 “mm.~ mmvo. mm.H m hmo¢o.o ummH mono. oH.H NHomo. NmoH m uHmmoo umo.H osmo. mo. ~ Ho.m uomé hmemo. ummo COHHDHom Guacamonm OH mm mxnmEmm «mamemcc H a mo oonuoz . H..mxm.x m . - m vogue: muHsouno oonHooz on» an oosoHHom no mHuoH mo .oooHooHom HH>V mumuumm mpoHHo mo coHuHmomsoooo .H> mHnme uoHEumuoHoE ucHsnoHosuH H-8H8 -¢0sz -on.oxo.x ..oxox HH>eoo mssm.oz 107 ‘4 «mHmhamcm mxumsom mo convex ho.HN m.NN o.om Hmm.oe o.- me. .o.- uom.ov .o.- Hom.ov em.~ Hoo.ov mo.m hHo.m Hmo.oe COHHDHom mumnflmonm m.OH mm mhvo. uommo.o memo. uoomo.o mmNo. um¢N0.0 coausaom wumconumu 0H ma H IGHEI ITHOE CHEIQH OEIH CHE H em“ I u H m xvsx muoax.mxm.x ..mxm ooocHuooo .H> loma mm.m a mooN uo¢.N uSYN m mh.H “mm.a N mm>.o H omw. N¢m¢.o N «m.N “MN.N N mm.H umm..n N HN.H “www.o N -¢OHXE AH>vmm mcsm.oz g. _...c. . u.. 1‘... w I N. am? 108 m ~e.o H¢~.V me.o comwwHom mumnOmono o.oH mg en..H m.H~ oH.o m mo.~ seuon~.m m m.H~ Hmmm.e ee.H zmuonm.e N m.~m Ho~.e oH.H zenon¢.~ N mom.o mm.o zvuono.H m oom.o ~m.o zeuono.H m om~.o m~.o zeuonH m m.o~ Hom.e mo.o zmnono.m m m.mH Ho~.ov mH.m HHHHVom zmuonm.¢ ucomouo HHHaHuHcH m o.oH HoH.oV mmv.o HHHHeomz ov.m H e.- .o.- Hmm.v .mm.m H ~.- xm~.ov He.H mxumaom mwmhamcd CHE: mHoE :«EuoHoEIH use - OHM: « HI wm HI oaxomxm .mem g mo oonuoz uxH. xm.x mu .x x HH>emm omocHuooo .H> oHnme mcsm.oz 109 smuono.e m ‘ m.om ucmmmum sHHmHuHoH AHHHVOm z oaXo.H N ¢| N N m.mm ho.mm m.om N “v.0m uo.vm N N.mm “m.bm N Oh.m “mm.m mxumEom mwmhamsd HIMMEI IOHOE HICHEIOHOEIH mo convex swa..mxm=x muoax.mxm.x UTSGHUGOU .H> magma Hem.v No.0 eon. em.o oom.o mm.e Hoe. mo.~ nee.v .eo.~ Hoe. mo.H nom.v uovH om.H Hom. u2H -mm.e .H~.H Hon. oo.H nem.v .mo.H oe.o Hom.e .oe.o H-oHs .oonz ..mxmx HH>vmm N mcsm.oz 110 COHuHm nomEoomo mswusc nouuuHmHooum AHHHVom S¢IOHXO.N Z¢IOHXN.H .ucommum sHHmHuHoH HHHHVom zmuono.e N mXumEom mflmhamc€ mo vogue: m5. mm.N umm.o hom.N hm. hom.N “05.0 «mm.H mm.o mm.o ooHuoHom wumnomono H.Hw mo 0.0 I 0.9m Nm.m mmN.o m.mN mNm. m.ma komm.o u00.0 uNmm. umv.v noom.o hmm.N 05¢.o mm.H om¢.o mh.o 0.0m Aamov fim.a HiawfilwanoE HacHEIwHoEIH HIdHE .fiomxz Ira..mxmxx m OHXnmxo.x ..mxmx AH>V0m @mficflvcoo .H> UHQMB a mcsm.oz 111 ( mXumEmm «mfimxamcm mo vogue: oauumfiouonm Hmv ouHsouso ooHMHooz .HV . N¢.o hNm.o Nomlmm. ooHuoHom oonouomm esHoom o.mH mm 00. “mm. om. “up. ooHuoHom monouon soHoom o.~H mm mm. «hm. om. u¢®.o uNh.o cowufiaom mondmosm soHoom m.HH mm om.o o.H kNo.0 “0.H “55.0 COHUSHom mungflm0£m m.HH mm HICHEIWIOHOE 1rd..mxmax HucHE|mHoEIH ,oax.mxm.x MI ooooHucoo .H> «Hume CHE HI! - ..mxox AhoN o.HHumb. “mo.o om.m uoo.m om.N hom...” mN.v uom.N ¢0.m “mm.N uGain H m mcsm.oz 11».E:5...|>.3a=1 J lmwj 112 plotting the log k against pH, a graph was obtained similar to that made in the previous section for 10"3 M ferrate solutions (Figure 17). A maximum rate of decomposition was found at approximately pH 11.5, after which the rate of decomposition decreased with higher pH. Comparison of values for the pH 9.5 to 11 range agreed with the conclusion arrived at in the previous section regarding a constant dependence on the hydroxide ion concentration. For dilute ferrate solutions, the decomposition rate increased slowly from pH 9.5 to 10.0, increased rapidly from 10.0 to 11.0 and then continued to increase to a maximum in the vicinity of pH 11 to 12. Thus, there does not appear to be a constant dependence on hydroxide ion for this pH range. The increased solubility of Fe (III) in the pH 13 to 2 M sodium hydroxide solutions compared to the pH 11 to 13 range, permitted an investigation of the decomposition of dilute ferrate solutions in a homogeneous system. Although the solubility of Fe (III) is not as high as in the pH 10.0 phosphate solution, it was sufficient to allow a moderate variation in the initial ferrate concentration. For some cases, the ferrate solution studied, remained homogeneous during the decomposition, but precipitation 112:3 EEK. H?! 55ml . ‘0 y u .u. 113 occurred on standing. The direct photometric method was used to follow most of the decompositions. The decomposition of ferrate in pH 13 - 2 M sodium hydroxide solutions was similar to the decompo- sition in pH 10.3 and 10.6 buffers. When graphs were made by plotting log A against time, they were similar to Figures 23 and 24. However, the apparent induction period was shorter for these solutions and appeared to decrease as the basicity increased. Figure 27 shows the decomposition in pH 13.4 solution. There also was a smaller variation in slopes for different initial ferrate concentrations, especially with the more concentrated sodium hydroxide solutions. Addition of ferrate (VI) to solutions initially containing Fe (III) again gave a linear log A plot with a higher rate of decomposition. But for these solutions, the magnitude of the increase was less than in the pH 10 - 10.6 solutions. Often the decomposition in a solution that initially contained Fe (III) was only slightly faster than the linear portion of decomposition that did not have Fe (III) initially present in the solution. When more than one decomposition run was followed in the same solution, it appeared that the decomposition rate became slower as the Fe (III) content built up in the solution. Investigation of this phenomenon was limited 114 FIGURE 27. Decomposition of Ferrate (VI) in pH 13.4 Sodium Hydroxide Solution as Followed by the Direct Photometric Method. 1.0 : Ao Fe(VI)x10‘4M Slope min.1 1 .585 1.97 -{138 T 2 .385 1.30 -.125 8 - 3 .148 0.50 -.0980 ,, ' 9 4 .178* 0.60 -.148 *Fe(VI) Added to Decomposed Solution 1 o 9 .Of 3 S 7 e a Q A 3 ‘ C 1 o 09 v 0 \5 0.1. .~ ‘ ‘ H. 6 o ‘9 d o _ . - 5 e ‘ I _k \‘ \ ‘ “ 028 V ‘ ' o — o 0 :0 1 ' 4 o 2 °°° 4. : 3 4 : 0 10 20 Minutes 115 as a result of the low solubility of Fe (III) in the solution, but decomposition in solutions discussed in a later section indicated this observation was correct. The rate equations which were developed for the decomposition of ferrate in pH 10.3 and 10.6 phosphate solutions also described the decomposition in pH 13 to 2 M sodium hydroxide solutions. It was found that the rate equation with a one-half- order dependence in Fe (III) gave linear plots with slightly lower initial ferrate concentrations than noted for the pH 10.3 and 10.6 buffers. This equation also described the decomposition up to an approximate ferrate concentration of 3 x 10-4 M, which was the highest concentration used for this pH range. Decomposition of ferrate in solutions initially containing small amounts of Fe (III), (5 x 10'5 M) was also described by the above equation. With higher Fe (III) concentrations, the decomposition appeared to be first-order in ferrate. The rate equation which was first-order in both ferrate and Fe (III) was used only for very dilute ferrate solutions (8 x 10-5 M). Linear plots similar to those shown in Figures 25 and 26 were obtained for decompositions in this pH range. Although fair 116 agreement is found between first—order rate constants for each solution, linear log A plots were only obtained after partial decomposition and it was felt that the above equation described the 1.3 to 1.9 M sodium hydroxide solutions more accurately. The decomposition of ferrate was studied in 1.25 M sodium hydroxide without sodium perchlorate in addition to the pH 13 - 2 M sodium hydroxide solutions maintained at an ionic strength of two. For dilute ferrate solutions, the decomposition rate compared very closely to decomposition in the 1.3 M sodium hydroxide solutions containing sodium perchlorate. Comparison of the decomposition rates in these two solutions for 10’3 M ferrate concentra- tions showed that the decomposition rate in 1.25 M sodium hydroxide without sodium perchlorate was slightly lower than in the 1.3 M sodium hydroxide solutions. Comparison of the decomposition rates found for 10'3 M ferrate solutions with those obtained from dilute ferrate solutions, indicated that a slightly higher decomposition rate was found in the dilute ferrate solutions. For both systems, the decomposi- tion rate decreased slightly with increasing alkalinity of the solution. The decrease was more noticeable 117 for the 10.3 M ferrate solutions and may have resulted from the increased solubility of Fe (III) in the more concentrated sodium hydroxide solutions. 3 M initial Although decompositions of 10- ferrate solutions gave linear first-order plots, it is probable that Fe (III) is also involved in the decomposition. When a solid Fe (III) phase was present, the amount of Fe (III) in solution would be expected to remain constant and the decomposition would then become pseudo-first- order in ferrate. However, the decomposition also appears to be first-order after a small amount of Fe (III) is present in solution, suggesting that Fe (III) accelerates the decomposition up to a point after which the rate is controlled by another factor. Data for the decomposition of ferrate in pH 13 to 2 M sodium hydroxide solutions are given in Table VII. For decompositions in solutions that initially contained Fe (III) and gave linear first-order plots, an experimental first-order rate constant is given. For data that fit the rate equations developed for the pH 10.3 and 10.6 solutions, the respective experimental rate constants are given along with the value calculated from the linear portion of log A plots. 118 ucmmoum MHHMHuHGH HHHHVom smuono.m wxumEmm o.¢m ¢.Nm m.om ho.Nm m.mN Hmm. mw.m hom.m Amm. Hem.v Hom.v .om.v Hem.o nvm.v ooHuoHom moHMouosm esHoom o.mH mg m. “Tom o.oH o.oo uo.oH. o.om Nmoo mm.o kom.® Hmm.w Amm.e Hom.v Ho~.. ooHpoHom monouomm soHoom o.mH mm HICHEIWIGHOEIWH .omxmrx m ICHEIHImHOElH IOonmxmx HICHE .mxmx vogue: oHuuosouogm uomuHo sh oo3oHHoo mm moHMouomm soHoom z m . MH mm mm.o HH.~ em.H nmmH we.o om. uom.o m~.~ m¢.H uomH om.o No. “3.0 OHMzHH>Vom msoausaom AH>V mumunmm vaHHQ mo coauamomEOUmQ .HH> magma mcsm.oz 4) 119 pcwmoum >HHMHDHCH HHHHvom ZmnOHMe udwmoum NHHmHuHCH HHHHVmM SmuoHMo.m mxumEOM 0.0m h0.0N u0.0N MNNM. m.0m 0.mm 00.0 e. M o.om mv.o 0.0N 0.0N H0.0m A00. 0.Hm mN.0 “00.0 Ham. 00.5 ooHuoHom oonouosm soHoom m.mH mm H.H + 0.Hm Hm.o o.Nm CHE: m 0&1 I I HI . m: H ma HIGHE HImHoE H ..mxw=x oax.mxm mu x omooHpcoo .HH> oHnme Hmmm. mm.~ .evm.e .me.~ .me.~ Hemm.v om.H Heom.0 Hm.o Hoo~.0 mo.o Hmm.v eo.o oom.o . mm.~ hom; me.H .mm.H Hmm.v oo.H .om.v oe.o .oo.o Hom.0 mm.o Hem.v me.o .mmmwm ¢0HNSAH>VQK H mcsm.oz 120 rilllih m.H H 82 pmumumeooum HHHHvom .ommo ooguoz muHEouno omeHooz mum xmom.o m.0m no.hm poms convex- .1: muHEonpo vOHMHUoz mm~.o mH.v smuone o.e~ imm~.v ~e.o m.o~ umSN Ammm. hawn; mm.m “mm.m m.mm “m.om Homm.xmo~.v mm.H umo.H o.om Hmo~.0 me.o o>.m HomH.ov 00.0 OHMHOHnoumm Eanom oz .moHMouomm eoHoom z m~.H o.H_H m.om Z¢IOme.H mmm.o No.0 ucmmoum >HHMHuch HHHHvom EmIOHxh omm.o m¢.o mxumEmm chfifilmlmaofilma HIGHEaHIQHoEIH HICHE OHXEAH>me N omuflmgx M'OHXOmUflmx e mxmx omooHucoo .HH> mHnme H mflflmooz 121 zouonm.H amuonm zonoHMom.H ucmnmum NHHMHDHCH AHHvam Emuonm mHN.o H.om HmH~.0 m.m~ .v.e~ loom..mem.0 e.om .o.m~ loam..mm~.0 om.o .o~.m AmmH.ioedv QOHusHom Oponun m Edspom 00.H o. No.3 oem. mom. e.em .m.om Hmmm. .oom.0 o.om .m.o~ Hmmm. .Hom.0 ~.e~ .m.m~ Hmmm..mom.0 00.m Ammmov COHusHow mponuvmm EsHoom oo.H mxumfimm cHEI mHoEI H H- ' I 0 w ..mxmmx H mIme.mxmx .Hmex omooHuooo .HH> 0Hhme CHE: mHoEIH CHE 0m.0 00.0 mm.N hMN.N m¢.H N«TH mo. “3.0 50.0 mm.H mn.~ “MN.N mm.N hmm.N om.H “0N.H mm.o 0Hx2HH>vwm w mcnm.oz 122 EvIOHX¢.N ucmmoum >HHMHuHcH HHHHVom zouonm.H mxumEmm o.H h m.m~ m.e CHEI 0 OE... CdEl 0 OE! 7.1.0 .m 7. 7W . mxm:x muon mxmx omooHHcoo .HH> meme 0NN.0 m0.0 NON.0 HH.H HICHE HV0HMSAH>me ..mxmx mcsm.oz 123 Studying the decomposition of dilute ferrate solutions was subject to several possible difficulties. The presence of trace amounts of oxidizable or catalytic impurities was one possibility. A second was that in using small amounts of solid ferrate, a slight amount of decomposition during handling would greatly alter the Fe (III) content and would effect the decomposition rate. Uniform sampling also becomes more difficult by this procedure, but it is felt that these problems were kept at a minimum. Decompositions followed by the direct photo- metric method were not stirred except during the initial mixing of the solid ferrate with the solution. Decompositions followed by the modified chromite method were stirred. Although comparable rate values were obtained for similar initial ferrate concentrations by the two methods, part of the variations in the decomposition rates obtained by the direct photometric method may have resulted from the solutions not being stirredvduring the decomposition. IV: ‘2:— - .._. 124 Decomposition ongilute Ferrate Solutions, 1 - 7 M Sodium Hydroxide Solutions The decomposition of potassium ferrate in l to 7 M sodium hydroxide solutions was followed by the direct photometric method. The solubility of Fe (III) in these solutions increased as the sodium hydroxide concentration was increased and was approXimately 8 x 10'4 M in 7 M sodium hydroxide. Ferrate 4 M were used for concentrations of 4 x 10'5 to 4 x 10' decomposition studies described in this section. In an attempt to maintain a constant ionic strength, the ionic strength of several sodium hydroxide solutions was adjusted to seven by adding sodium perchlorate. The ionic strength of 7 M sodium hydroxide was assumed to be seven. This procedure served only as an approximation as the activity coefficients of sodium hydroxide and sodium perchlorate are different at these concentrations (32). In addition to the above solutions, sodium hydroxide solutions that did not contain sodium perchlorate were used and except for the 7 M sodium hydroxide solution, these solutions will be so designated. The effect of Fe (III) on the decomposition rate was studied by the same procedure as used for dilute ferrate solutions described in the 125 previous section. The decomposition rate of ferrate in the pH 10.3 - 10.6 and pH 13 to 2 M sodium hydroxide solutions has been shown to be dependent on the Fe (III) concentration. As the concentration of Fe (III) increased, the rate increased and then remained approximately the same. There was some indication that the rate decreased slightly at higher concentrations of Fe (III). For the more concentrated sodium hydroxide solutions used in this section, the decomposition rate was found to decrease as the Fe (III) concentration in solution became higher. With the 2.5 and 4.0 M sodium hydroxide solutions that did not contain sodium perchlorate, the decompo- sition rate of ferrate appeared to be dependent on Fe (III) until the Fe (III) concentration reached a particular level, after which the decomposition rate appeared first-order in ferrate alone. In the 2.5 M sodium hydroxide solutions, the decomposition rate of 5 x 10’5 to 1.4 x 10'4 M ferrate solutions was described as being first-order in ferrate and one-half-order in Fe (III) and except for the S x 10"5 M ferrate solution, when the log A was plotted against time, the linear ‘ portion of the curves had approximately the same slope. For a 3 x 10"4 M ferrate solution or for ferrate solutions that initially contained Fe (III), the 126 decomposition was first-order in ferrate and the slopes were comparable to the slopes obtained from the linear portion of log A plots for the more dilute ferrate solutions. In the 4.0 M sodium hydroxide solutions, decomposition of 7 x 10'5 to 3 x 10'4 M ferrate solutions appeared to be first-order in ferrate, although the early portion of the more dilute ferrate solutions deviated from the linear log A plot. For the 2.5 and 4.0 M sodium hydroxide solutions, the only indication of a slower decomposition rate as the Fe (III) concentration increased was when the Fe (III) concentration approached the solubility limit of Fe (III) in the 4 M sodium hydroxide solution. Decomposition of ferrate in the 2, 2.5, 3, 4, 5, 6 M sodium hydroxide solutions that contained sodium perchlorate and the 7 M sodium hydroxide solution was similar although the rate of decomposition in the different solutions varied. Linear plots were obtained for the more dilute solutions when log A was plotted against time. Only the very early portion of the decomposition showed an induction period and the slopes of the plots were similar. As the initial ferrate concentration was increased, log A plots were no longer linear for the last portion of the decomposition and for 3 - 4 x 10'4 M ferrate solutions, 127 curves were obtained for the plots and the decomposition rate was slower. Successive additions of ferrate to the same solution showed that as the Fe (III) concentra- tion increased, the decomposition rate decreased. The behavior is shown in Figures 28 and 29 for 7 M sodium hydroxide solution and shows the typical decomposition pattern in these solutions. If it is assumed that the decomposition rate is inhibited by Fe (III), an expression can be developed for the decomposition rate, using the initial ferrate concentration and the amount of ferrate remaining. Thus, -dI§e(VIfl = kml'FeLvr dt {Fe(III) . Substitution of c for [sewn] and (Co-C) for [Fe(III)] and integrating (37) gave C —-ColnC = k"° t + constant or (c -- co) -- Coln C/Co : k" t for the equation. The most convenient way of plotting the data was to use the absorbance of ferrate directly and plotA-AolnAagainst time. The experimental rate constant was Obtained from the slope of the resulting straight line by converting the slope into terms of moles of ferrate per liter. Typical plots are shown in Figure 30, curves A and B for 7 M sodium hydroxide. Curve C shows a plot for a solution that initially 128 FIGURE 28. Decomposition of Ferrate (VI) in 7 M Sodium Hydroxide Solution as Followed by the Direct PhOtometric Method. M Fe(VI) x 104 -Slope min"1 1.0 W A 0.77 0.0565 B 1.37 0.0625 0 C 2.19 0.0556 D 3.63 (0.06 initial) 1 n 0 \\ e e o o 0 ' . o 9 e _ 2‘ o , ‘ ¢ 0 e ’ I O C . D .' - 0.0 } :7A i g g 0 40 80 Minutes 129 FIGURE 29. Decomposition of Successive Portions of Ferrate in 7 M Sodium Hydroxide Solutions. 1.0 —" M Fe(III)x104 M Fe(VI)x104 -Slope min“ : 1 1.68 0.0583 2 1.68 1.4 (0.03) - 3 3.1 2.0 (0.01) -f o a o 6 .J. . 0 up \ \‘ u e 2 \. \ \“ 1 0.0 % g i 1 0 40 80 Minutes 1.7 130 FIGURE 30. Decomposition of Ferrate (VI) in 7 M Sodium Hydroxide as Followed by the Direct Photometric Method - kuexp. plot. 3 '02 . * M FeXIII) M Fe VI) 6 ' Slope~ x 10 x 10 _ o 1 0.0570 -—- 2.19 ” 2 0.0555 .1“- 3.64 a " 3 0.0557 3.8 0.98 O * Absorbance Units - min-1 ’I . I 0‘. 2’8 i i .L l i ! 0 20 Minutes 40 60 131 contained 4 x 10'4 M Fe (III). The curvature of the early portion of curve A is characteristic of the more dilute ferrate solutions that are described by this relationship and is probably a result of an insufficient amount of Fe (III) to cause a reduction in the decomposition rate. When Fe (III) is initially present, there was no curvature at the beginning of the plots. For solutions that initially contain Fe (III), a C6 value must be used, where C0 was the sum of the initial Fe (III) and initial ferrate concentrations. The initial ferrate concentration for each was again estimated from the initial absorbance of the ferrate solution. However, with solutions that initially contained Fe (III), the value was not readily attainable due to an interesting phenomenon that occurred on addition of the ferrate. When ferrate was mixed with a solution containing moderate or larger amounts of soluble Fe (III), a rapid decomposition occurred with the evolution of gas bubbles that lasted for only a minute or two, after which the solution cleared and a dilute ferrate solution remained. The ferrate solution then decomposed similarly to the normal decomposition except at a slower rate. Thus, the initial ferrate concentration had to be estimated 132 by extrapolation of the decomposition data and this introduced some uncertainty of the value. This initial decomposition with solutions that contained Fe (III) was more pronounced for the more concentrated sodium hydroxide solutions, but was found in the less concentrated sodium hydroxide solutions which had a higher Fe (III) content. For very dilute ferrate concentrations (4 - 7 x 10‘5 M) and in 2 and 4 M sodium hydroxide solutions without sodium perchlorate, the decomposi- tion rate can be described by the rate equation which assumes a one-half-order in Fe (III) and a first-order in ferrate; Slopes from the linear portion of log A plots were slightly lower for these ferrate solutions than for the higher ferrate concentrations. For the high ionic strength solutions, the ferrate concentration at which a reduction in the decomposition rate occurred, varied for the different solutions, but appeared to be approximately 1 - 2 x 10.4 M in ferrate. Figure 28, curve C, shows that for 2 x 10-4 M ferrate solution, the decomposition appears to be first-order in ferrate for a large portion of the run and extrapolation of the linear portion gave a slope similar to the more dilute solutions. The same decomposition data are plotted in Figure 30, curve A, 133 and shows that the inhibition by Fe (III) occurs after a quantity of Fe (III) is present. The comparison of rates in solutions with the same sodium hydroxide content, with and without sodium perchlorate, showed them to be similar. ith 4 M sodium hydroxide a slightly lower rate was found for the solution without sodium perchlorate, but the difference is not considered significant. A comparison of ferrate decomposition in 1.25 M sodium hydroxide solutions that contained different amounts of sodium perchlorate, indiCated that for solutions with a high sodium perchlorate content, the decomposition was inhibited by Fe (III). Decomposi- tion of ferrate in 1.25 M sodium hydroxide without. sodium perchlorate described in the previous section did not indicate a reduction in the decBmposition rate as the Fe (III) concentration increased. However, the decomposition rate of l x 10"4 M ferrate solutions that did not initially contain Fe (III) were almost identical in the solutions with different ionic strengths.“ Comparison of decompositions in 2.5 M sodium hydroxide with and without sodium perchlorate showed a similar trend. Decomposition in sodium hydroxide solutions with a high ionic strength in general resemble the decomposition of ferrate in 7 M sodium hydroxide. 134 Comparison of the decomposition rates in solutions containing sodium perchlorate showed that the rate decreased as the sodium hydroxide concentra- tion was increased. A plot of log kex against p. log (0H) for first-order rate values and k"' values showed a linear relationship. There was considerable scatter in the first-order values, but better agreement among kgkp. values. The slope of the resulting straight line suggests that hydroxide ion inhibits the decomposition and by a factor of (OH)%. Table VIII gives the experimental rate constants and ferrate concentrations for decomposition in the solutions discussed in this section. Several of the first-order rate values were obtained from the most linear portion of log A plots and are listed only for comparison purposes. 135 ucmmmum hHHMHuHCH AHHvam zvuOme.~ om.m unemoum mHHMHuHCH AHHHmeHWIOwa NH.m Nn.m mm.m nmh.m m0.m HeH.0 HeH.o0 16H.o0 hNN. uMNN.0 mNN. k0MN.0 h numcmuum UHGOH .mUonupzm EsHUom 0m.N 00N. 5.0m "NmN.0 h Suwcmhum UHGOH .mnfixoucwm Edflfiom S H0.N mxumfimm I. H0 l0 OE I I HuoHe H- ummmxw HuoHe mHmHos xH .H..oHs mon :x .oxo.x .oxmx .oonumz OHuumEouonm uoman on» ma pm3oHHom mm monouosm soHoom cH z e u H .oooHuoHom mumuumm ousHHQ mo COHnHmomEooon .HHH> THQMB Hm.H Hm.0 0N.m 0m.N “Hm.N Nm.H “N0.0 0m.H ¢0Hx2 AH>vmm mcsm.oz 136 ZfilOHXmoH ZfilOHXm ZVIOHXmoN ucmmmum hHHMHuHcH HHHHomm zeuon¢.H usmmoum aHHMHuHCH HHHHme zvuono.e _mxumEmm m0N. n.0N “mmH.0 5 numcumum UHGOHiioUHxOupmm Edeom 2 m0.¢ H acHEI mo. H.HQH moo._H m-.o HH.m HoH.0 ~o.m HmH~.0 eo.m Hmo.0 om.m HmH.0 mo.m hmo.m HNH.0 oo.~ o-.o mH~.o .¢m~.o mH~.o hm~~.o mm. H 38 moo. H 626 e0.m Heo.0 HnumuHHlmHoE anHEImanofilmH Ilmnmmm. mon..mmwx ..mxm=x ..mxmx omooHucoo .HHH> mHnme N0.H uM5. N H5.0 H 55.0 H 0N.H H m0.H H H0.N N0m.N N 00.H H 00.H hmm.H N 05.0 “05.0 N N0.N H 00Hx= mcsm.oz HH>00m 137 00H.0 mmoH h05H. h0m.H h¢0H.0 u0N.H m 55H.0 «0.0 H 5 numcwuww UHGOH .mmMXouUNm Echom 2 m0.m eH..H om.~ moo. u eo~.o :e-onm.H oe.~ Hmo.0 oo.~ H Zv-onH.H os.~ Hmo.ov om.~ H zv-onm.H mo.m Hmo.ov oH.m H ucmmmum mHHMHuHcH HHHHVom Zm-onH.¢ MH~.o me.o . H oH.m oo.m ho~.m HHH.V .mo.m m OHN. oo.H mm.~ .me.~ .mo~.o nmmH m mom. Hm.H uoH~.o .mo.H m m HmEm CHE- um H no oE :HE- w oEI CHE x2 mcsm.oz x m H- H- u H H H- m- H mH H- eoH mon..mxw=x ..mxm:x ..mxmx HH>vmm ooooHuooo .HHH> oHnme 138 Se-onm.m zv-onm ze-on5.~ ze-on¢.H ucwmoum >HHaHuHcH HHHHVOE :m-onm anmEmm mN.N nVH.N voH. h00H.0 0¢H.0 0mH. ummH.0 5 numcmuum UHGOH .mUonup>m EsHpom z mm.m 00.H u00H HN.H 0H.H "m0.H 00.N 55.H mN.H 0¢.H 0N.H 00.N “00.N 00.H u00H N NH..H oo.H moo..u omH.o om.~ xeo.0 m¢.~ Hoo.0 m¢.~ H5o.0 o5.~ HeH.0 Hm.~ 155H.0 HovH. ov.~ kHH.H “o5H.o0 HmvH. om.m .65.~ uomH.o0 HIGHE-H-uuuHHlmHoE ImHoE-mH HICHE m0Hx..me= ..mxm=x ..mxox ooocHucoo .HHH> mHnme 0H3: w AH>vom mcsm.oz 139 ZmIOme Z¢I0Hxv an EVIOH H ze-one ucmmmum 5HHMHuHCH HHHHomm ze-one.H ummsHH uo: uoHO a moH mxumEmm 00.H 50.H 00.H hN0.H h0v.H u00H AHH.V 0m.H HNH.0 eo.m ¢mH. 0N.N k5NH. h0H.N vMH. m0.H u0NH. kmm.H k33.0 u5m.H mNH. M5.0 homH.0 h55.0 #QUHXOH053 EdHnom 2 00.0 0H. + Hm.N 0N.N H¢.N 0m.N 5¢.N N¢.N H000...._... 00H. H-cHE-H-umuHHlmHoE m H OHx s n mxma-ox UmscHusou CHE... 0 OE... I . wml H ”H ..mxm.x .HHH> QHQME Hmo.0 oo.H Hoo.0 mo.H Hmo.0 HH.H H00.0 0m.m mm.m --m-m_ - t H- H eoH ..mxwx HH>me H upsm.oz 140 w.o.M_H.mm oHoafl mmw.o sv-onm.~ on. em.H H zv-one.H , mum. mm.o H oem. mm.~ .me~.o .mm.~ N oem. H¢.H ~.e~ .o.6~ .mm~.o hmmH m m.om HH5H.0 vo.o H « wuonuwhm Edeom E 00.N 6H. H~5.H mo. HHMHo ze-Ome.m em.H Heo.0 mo.o H zv-OHx5.H mm.H “50.0 N¢.H H ZvIOHXN.~ 00.H A50.V 00.N H ucmmmum 5HHmHuHsH HHHHme Zeaonm.N m5.H Hmo.v 0N.H H m HmEm CHE: nmuH um OE sHE- 0 oE- CHE x2 mcsm.o x H- HI H H H- x- H mH _H- «0H 2 - mon..mxm ..mxm ..mxwx HH>vmm - a: :v— omocHuooo .HHH> mHhme 141 1 I! COHusHom on» on 0000M uoc oHMMOHnouwm Echom s ucmmmum sHHnHuHoH HHHHVom Zm-onm.H.o.5 u 39H ommo sumcmuum UHCOH N5.¢ AO0N.v 00.H H mmN. 05.H u00N.0 u00.H N 0.0 u NmN.0 m0.H numcmnpm UHCOH -III“5.0N ummN.0 10m.H N 00Hx0u0>m EsHpom z mN.H COHuHm 0.0m N00. I 00H.0 Iomaoomp mo 0:0 Mafia H030Hm fluflm .va . 0 mm. o N H HmH. N5.H N00H.0 um0.H N 00H. 0N.H 0.MN ko.mw u50H.0 h05.0 N 4 monomvnm esHoom 2 No.6 mxumEmm IcHEIHIHmuHHImHoE H IsHEIxImHoEI H HICHE OHM: mcsm .02 H m0Hx..mxm ..mxmmx ..mxmx HH>00h 0055050 .5: 03,3. 142 Rate Studiesiin l - 7 M Sodium Hydroxide Solutions, 10'3 - 10"2 M Ferrate (VI) The decomposition of ferrate in l - 7 M sodium hydroxide solutions was followed using the modified chromite and oxygen evolution methods. 3 to 3 x 10‘2 M Ferrate concentrations from 1 x 10- were used, which resulted in formation of a Fe (III) solid phase during the decomposition. The decomposition of ferrate in these solutions was originally followed using the oxygen evolution method. The modified chromite method was used to check the decomposition rates in 1 to 7 M sodium hydroxide solutions, when different rates were obtained by the two methods in pH 7.7 to 14 solutions. For the more concentrated sodium hydroxide solutions, the decomposition rates obtained by each method were similar. Only for a few solutions in which the. decomposition rate was more rapid, did the two methods give different values for the decomposition rate. Decomposition rates measured in this section were in general much slower than measured for the pH 7.7 - 14 range and the oxygen evolution method apparently is satisfactory for the slower decomposition rates. Decompositions followed by the modified chromite method gave linear plots when the log of the ferrate concentration was plotted against time. Several of 143 the decompositions were only followed for two half- lives because of the slow decomposition rate. Decompositions followed by the oxygen evolution method gave linear first-order plots, but many of the plots deviated after two half-lives had passed. The deviation was similar to that found in the pH 7.7 - 14 range for the oxygen evolution method and it is felt that the deviation was not a change in the decomposition rate. This is supported by the fact that if the solution was not stirred, the deviation occurred earlier and was more pronounced. Decomposition rates were obtained from the linear portion of the oxygen evolution plots. Decomposition of ferrate in 1 - 7 M sodium hydroxide solutions is first-order in ferrate and plotting the log of the ferrate concentration against time gives a linear relationship. It is very likely that Fe (III) was involved in the decomposition of ferrate in these solutions, but with the higher ferrate concentrations, the solutions became saturated with Fe (III) very early in the decomposition and the decomposition became pseudo-first-order. With the more dilute ferrate solutions, plots of the decomposition data showed a curved section immediately following dissolution of the solid ferrate indicating that the 144 decomposition rate was slowing down as the concentration of Fe (III) was increasing. The remaining part of the plot was then linear. Comparison of the decomposition rates from 1 x 10'4 M ferrate solutions with the rate from 10'2 M ferrate solutions shows that the latter solutions decompose at a much slower rate, indicating that the decomposition was inhibited by the Fe (III) in solution. The first-order decomposition rates for the more concentrated ferrate solutions did not change as the initial ferrate concentration was varied, as expected if the Fe (III) concentration in solution remained constant. For 10‘3 - 10‘2 M ferrate solutions, decomposition rates were lower in the more concentrated sodium hydroxide solutions. For solutions with the same sodium hydroxide content, a lower rate was found for the solution with a high ionic strength. For the dilute ferrate solutions in the previous section, a lower decomposition rate was also found for the high sodium hydroxide concentrations. However, the difference between the decomposition rates in 1 M (without sodium perchlorate) and 7 M sodium hydroxide solutions, was much greater for the 10'3 - 10'2 M ferrate solutions than for the l x 10"4 M ferrate solutions. This was a result of the inhibition of Fe (III) in the 7 M 145 sodium hydroxide solution due to the higher Fe (III) concentrations found with the 10‘3 - 10‘2 M ferrate solutions. For the l x 10"4 M ferrate solutions, the decomposition rate in 7 M sodium hydroxide was approximately one-half that in 1.2 M sodium hydroxide (without sodium perchlorate). For the 10‘3 - 10’2 M ferrate concentrations, decomposition rates in the above solutions, differed by factor of about 100. When the log kexp. was plotted against the log (OH) for decompositions in high ionic solutions, the best straight line had a slope of -1.95. This suggested that the decomposition rate was inhibited by (OH)2, but as previously mentioned, the ionic strengths of the solutions were not exactly the same. However, decomposition in these solutions showed that for 10‘3 - 10'“2 M ferrate solutions, the decomposition rate decreased as the ionic strength or sodium hydroxide concentration increased. Addition of hydrous ferric oxide to 1.2 or 2.5 M sodium hydroxide solutions of ferrate had no effect on the decomposition rate. Addition of hydrous ferric oxide to 7 M sodium hydroxide or to 3 M sodium hydroxide that contained sodium perchlorate, doubled the decomposition rate. The increased decomposition rate was not expected, as previous experiments indicated 146 that soluble Fe (III) retarded the decomposition in solutions of the same composition. It was felt that the solid Fe (III) would simulate the conditions found in ferrate solutions,following precipitation of the Fe (III) formed in the decomposition. Since the Fe (III) was present mainly as a solid phase, the increased decomposition rate was probably a result of a heterogeneous reaction, either with ferrate (VI) or some intermediate formed in the decomposition. This is also suggested by the fact that the same decomposi- tion rate was obtained over a fairly wide initial ferrate concentration range and that an increased decomposition rate was found only when approximately four times the amount of Fe (III) normally present at the end of the decomposition was initially added. The increased decomposition rate could be ascribed to impurities which had been added along with the hydrous ferric oxide, but it is felt that this was not the case. If it is assumed that the same mechanism is operative in the 10'3 - 10"2 M ferrate solutions as found for the dilute ferrate solutions, then it should be possible to calculate the experimental rate constant for the more concentrated ferrate solutions. By assuming that small amounts of Fe (III) solid formed in the decomposition did not alter the .th_ 147 decomposition rate, using the experimentally determined Fe (III) solubility and the rate constants found for the decomposition in homo- geneous solutions, an estimate of the decomposition rate in the 10'3 - 10"2 M ferrate solutions was made. Values calculated in this way were faster than the experimentally measured rate constants by as much as a factor of ten. This suggested that either the same mechanism.was not operative or that the solubility of Fe (III) in the ferrate solution was greater than that found for decomposed ferrate solutions. For the 7 M sodium hydroxide, only if the solubility of Fe (III) was assumed to be 10-2 M instead of the experimentally estimated value of 8 x 10'4 M, did the calculated and experimental values agree. However, lower experimental rate constants were also found in heterogeneous decompositions for solutions where the homogeneous decomposition did not appear to be inhibited by Fe (III). Experimental rate constants, ferrate concentrations and additional information are listed in Table IX and X for decompositions in l - 7 M sodium hydroxide solutions. 148 .COHusHom oumuuHc UHuuom 0oumuucoocoo mm 00000 AHHvam 0E mH “Hammoum hHHaHuHCH unenm 0HHom HHHHVom .0E 00 .00000 00on UHnumw msou0hm .mxmx mom OQOHm ummcHH 00m: “00>Hso HOHm mo :OHuuom HMHuHcH mxumEmm mH.e oo.o Ho.~ umo.~ m5.m .~.oH mo.H hmo.~ H.HH .~.5H om.e .m5.¢ H.5H .o.~H m.HH om.m om.e .5m.¢ 0.5H “8.0 00.6 ~.5H H~.o om.m 5.6m .m.~o 5.0H “mm.m 5 .suocouum oHcoH .Hmmws mOszxH>00m OHM x m vogue: muHsouno omHuHooz 0:0 an omonHom mo moaxouosm H 00.0 N 00.0 N 00.0 N 00.N H 00.N N 0N.H mcsm.oz 00onu0mm soHoom muHumHoz EOH0om z 5 I H cH mumnumm mo COHuHmomEoomQ .xH «Hams 149 o .sumomuum 0HcoH m0 0 .nu0amuum UHCOH NOH m.¢H m.mH nuwcwuum UH:OH mchmm> me3 nCOHusHom o.HH uH.HH v.oH .68 mm .sHHoHuHoH 00000 00on UHuuum msou0>m mNH mNH ummH m.~H ohm .o5w om.m “00.0 05.0 “0H.¢ u05.0 mHMHOHzouwmIEOH0om usonuH3 mcoHHSHom .68 oo .aHHmHuHoH ouoou oono oHuuou uncut»: o5.a m~.o haemawm -mnst monzHH>eom mon..mxmx ooooHuooo .xH mHnoe H mnsm.oz 0N.H 0N.H No.0 00.N 00.N 0N.H 00.0 m0onu0Nm .EOH0om .muHumHoz 150 .mxox now 00m: EOHpuom HmmsHH “00>uso uOHm mo COHuuom umuHm mxumEom 0N.N mm.m 00.N h00.N 5.5H h0.5H v.0 h0.m 0.5H um.5H 0.0 u0.m 0.5H Nm.5H 0.0 hm.m 0.5H u5.5H m.m "0.0 0.5H h00.4.” 0.0H h0.0H 0.5H umv.m mm 10m 5.5H u0¢.m 5.5H 00 “mm uNm um.5H h00.N H5 “05 0.5H u05.N 5 .nuwcouum OHGOH HIsHE 0szHH>0mm 0Hx..mxmx m 0onvmz EOHusHo>m cmmhxo 0:» an 0030HHow mm mcsm.oz 00onn0hm E3H0om £5I H pH mpmuumm mo COHunomEoomQ .x mHnme 00.0 00.0 N0.¢ 00.m 00.N 00.N N0.N 0¢.H 0N.H 00.H 00onu0>m EsH0om huHHMHoz 151 .npmcuuum UHCOH .numcouum OHCOH .numcouum UHcOH .numcouum UHsOH m0 N0 50H 5NH oH.¢ .m~.v 00H 00H h50H uN0H 0.5H 0.5H 5.5H 0.5H um.5H hmv.m u00.m ouMHOHnouwQ E5H0om usonuH3 mCOHusHom mxumEmm oo.H “oo.H CHE H.‘ o mon oxmx 0.5H u0.5H OHXZAMVumm omooHucoo .x «Hams H 0N.H H 0N.H H 0N.H N 00.0 N 00.H N 0N.H N 00.0 mcsm.oz 00onu0mm EOH0om muHumHoz 152 Decomposition of Dilute Ferrate Solutions using a Galvanic Oxygen Analyzer. The decomposition of dilute ferrate solutions was also studied by following the oxygen generated in the ferrate solution with a galvanic oxygen analyzer. The oxygen analyzer is designed for low oxygen concentrations and ferrate concentrations similar to those used in the direct photometric method produced sufficient oxygen to give a satisfactory instrument response. The electrode for the galvanic oxygen analyzer was of single-piece construction and enclosed in a polyethylene membrane, which was permeable to oxygen. The most satisfactory arrangement was to place the electrode in a ferrate solution sealed from the atmosphere and record the instrument's signal until the ferrate had completely decomposed. Solutions used for the decomposition runs were purged with nitrogen for several minutes prior to the addition of the solid ferrate, to lower the oxygen content of the solution. The dataIwie plotted in the form, log (responseoo- response t) against time. The linearity of the galvanic oxygen analyzer's response was checked by measuring aqueous solutions containing varying amounts of dissolved oxygen. These solutions were made by mixing aliquots from two stock solutions, one saturated with oxygen and the other which was continuously purged with nitrogen. A linear response was found up to the maximum response of the instrument. The effect of the generation rate of oxygen on the instrument‘s response was determined by generating oxygen by electrolysis in a dilute perchloric acid solution, using two platinum electrodes in an H-type cell. The two arms of the cell were separated by a fine-sintered-glass frit and the galvanic oxygen electrode was sealed in the arm containing the platinum anode. Oxygen was generated at several different rates, which corresponded to the rates encountered in ferrate decompositions. It was found that the instrument°s response was not linear with the different generating rates, the response falling off with the faster rates. Apparently the electrode did not establish equilibrium with the solution rapidly enough to follow all of the generating rates used, even though the solutions were rapidly stirred. This observation is in agreement with the manufactor°s suggestion to rapidly stir the solution being measured, so as to obtain the maximum response from the instrument. 154 Although the galvanic oxygen analyzer could have been adapted for some of the ferrate decompositions studied, the modified chromite and direct photometric methods were preferred. Decomposition of Solid Potassium Ferrate (VI) at Elevated Temperatures. A knowledge of the chemical reactions of Fe (IV) and Fe (V) in aqueous solutions is important in understanding the decomposition of ferrate (VI). Solid compounds of iron in these oxidation states have been prepared and the compounds are reported to disproportion- ate in water or dilute sodium hydroxide solutions (23,24,25). In order to further study the reactions of these oxidization states, an attempt was made to prepare some Fe (V) and Fe (IV) compounds. The compound, K3Fe04, has been prepared by heating potassium and iron oxides at 450° in an oxygen atmosphere (25). Attempts by this author to prepare the compound by the above method were not successful. In an alternate approach to prepare an intermediate oxidation state of iron, the decomposi- tion of high purity potassium ferrate (VI) at elevated temperatures was investigated. 155 The thermal decomposition of potassium ferrate (VI) has been previously described by Scholder g; 31. (36). It was reported that between 2000 and 350°, a ‘black decomposition product was obtained which became partly coated with green KFe02 at 600°. At 800°, the active oxygen content was less than one active oxygen per iron and the solid product was a mixture, ‘.--—-H consisting of K3FeO4 and KFeOz. In the present study, decomposition of solid potassium ferrate was carried out in porcelain and platinum boats in a closed system provided with a mercury manometer. In decompositions carried out under 400°, it was found that the porcelain boats were being corroded by an oxide of potassium. With several decompositions, a white or light colored film was found along the sides of the platinum boat and suggested that an oxide of potassium was formed. Decomposition of the potassium ferrate appeared to begin at temperatures just above 120°, although the decomposition rate was slow. Introduction of solid potassium ferrate into the reaction chamber which was preheated to 300°, resulted in a fairly rapid decomposition and produced a black solid that appeared to be homogeneous. A weight loss of approximately 4% was found. Heating the potassium ferrate (VI) 156 above or below 300° or reheating the black solid higher than 300°, gave an olive green product with a 6 - 9% ‘weight loss. The olive green product was stable to over 450°, but continued heating eventually resulted in an Fe (III) oxide being formed. The black residue contained approximately two active oxygens per iron, while the green residue contained only small amounts of oxidizing power, which varied from run to run. When the green residue was treated with 95% ethyl alcohol, no apparent reaction took place, but the alcohol solution became basic. When the black residue was treated with 95% ethyl alcohol, a vigorous reaction occurred with the evolution of a gas. A green solid similar to the above mentioned solid was obtained and the alcohol solution was found to be basic. The percentage of iron in the green solid obtained from the alcohol solutions corresponded to that in the empirical formula, KFeOZ. Addition of the black solid to water resulted in a Fe (III) solid phase and a ferrate (VI) solution, as evidenced by the appearance of the red Fe (VI) color. The green residue, when treated with water, formed a Fe (III) solid phase. It is possible that the two products obtained from the decomposition of potassium ferrate at 157 different temperatures were the result of kinetically controlled reactions. When the black product was obtained at 300°, the pressure inside the reaction tube increased rapidly and then increased slowly with continued heating. At temperatures above or below 300°, the pressure increase was more uniform. At temperatures close to 300°, a definite mixture of green and black residues was obtained, which eventually became entirely green with continued heating. A satisfactory conclusion was not reached regarding these decomposition products. Analysis of the products was difficult as the residues were sensitive to moisture. For the present work, the objective was to prepare compounds of iron in an intermediate oxidation state in order to compare their reactions with those of ferrate (VI). The only product that possibly contained iron in an intermediate oxidation state decomposed immediately in aqueous solutions. It is possible that the black residue obtained in this work contained the black KFeO3 which has been reported (25). If KFeO3 was obtained, then the work suggests that an oxide Of potassium was also present in the solid residue. 158 Isotopic Exchange of Fe (VI) and Fe (III) The inhibiting effect of Fe (III) found in several of the ferrate solutions suggested that Fe (III) in the solutions may have been oxidized to Fe (VI) during the decomposition. Exchange has been shown to occur between Cr (VI) and Cr (III) in solution (38). In an attempt to experimentally determine if an exchange of oxidation states occurred in the iron system, a Fe59(III) tracer solution was made by dissolving a radioactive ferric hydroxide precipitate in 7 M sodium hydroxide. Potassium ferrate (VI) was dissolved in the tracer solution (maintained at 40°) and aliquots of the ferrate solution were periodically removed. Each aliquot was centrifuged and approximately 80% of the solution was placed in a second centrifuge tube, where a barium hydroxide solution was added to precipitate barium ferrate. After washing the precipitate (twice with sodium hydroxide and then twice with water), the precipitate was filtered on a small pre-weighed filter disc and washed with absolute alcohol. After vacuum drying, the precipitate was weighed. The precipitate was dissolved in hydrochloric acid, transferred to a counting vial and counted in a gamma scintillation counter. 159 Barium ferrate precipitates obtained from a series of samples taken immediately after mixing of the ferrate solutions, contained considerable activity. Precipitation of barium chromate from a fresh portion of the tracer solution using a similar procedure, indicated that the activity was due to contamination of the precipitate with Fe (III). When barium carbonate was precipitated from the tracer solution, the counting rate was only slightly above background. Contamination of barium chromate or barium ferrate with Fe (III) is not unexpected. The quantitative analysis of sulfate ion by precipita- tion as barium sulfate is known to coprecipitate Fe (III), (39). Purification of the barium ferrate by reprecipitation is not possible as the precipitate is insoluble except in acids, which cause rapid decomposition of ferrate. In an attempt to reduce the contamination found with the above procedure, an additional step was performed. After centrifuging the aliquot taken from the ferrate solution, lanthanum hydroxide was precipitated by adding lanthanum nitrate to attempt to lower the Fe (III) content before precipitating barium ferrate. After centrifuging, a second lanthanum precipitation was made. The solution was then used for the barium precipitation. H.o 160 Even with this additional step, barium ferrate precipitates obtained from aliquots taken shortly after dissolution of the potassium ferrate, still contained activity. However, by standardizing the experimental procedures, relatively consistent values were obtained for initial ferrate aliquots. When aliquots were taken over a period of several hours, the specific activity of the barium ferrate precipitates increased, reached a maximum value and then decreased in a uniform manner. If exchange did take place, the decrease in the specific activity of barium ferrate would have resulted from the lower ferrate (VI) concentration and the increase in the total amount of Fe (III) present. The specific activity of precipitates formed from aliquots taken shortly after dissolving the solid ferrate was approximately 25% of the maximum values obtained. For aliquots treated with lanthanum, separation of the final barium ferrate precipitate was complete approximately 10 minutes after removal of the aliquot. The maximum specific activity was found in aliquots taken about 70 minutes after the initial mixing of a 0.01 M ferrate solution. The experiments indicate that exchange might have taken place, but the separation procedure above 161 used was not completely satisfactory. One other separation method was tried in an effort to separate ferrate (VI) from the Fe (III) in solution, assuming that the Fe (III) species was FeOE. Tetraphenyl- arsonium chloride has previously been used (40,41,42) to separate MnOZ from.MnOZ using an extraction with chloroform. Attempts to use this material here were not satisfactory as considerable decomposition of the ferrate (VI) occurred, possibly due to the oxidation of chloride ion. Discussion of the Previous Studies on Decomposition of Potassium Ferrate (VI). One of the earlier studies on ferrate (VI) decomposition was done by Hinsvark (14) who studied the decomposition in 4 to 9N sodium hydroxide at 30°. 4 - l x 10'3 M ferrate The decomposition of 5 x 10- solutions was followed by removing aliquots of a ferrate solution and determining the absorbance of ferrate (VI) spectrophotometrically. He found that the rate decreased as the decomposition proceeded and used the initial slopes of log A plots for calculating rate constants. First-order rate constants found in this study at 400 were approximately one and a half times higher than the values reported by Hinsvark for solutions of similar sodium hydroxide content. 162 Hinsvark reported that glass wool only slightly effected the decomposition rate, but that light in the visible region accelerated the decomposition. However, the present author and others (3,22) found no effect by light of the visible spectrum. Hinsvark suggested that hydroxyl radicals might be involved in the decomposition. Magee (22) studied the decomposition of ferrate (VI) at 250 in solutions from pH 9.2 to 7 m sodium hydroxide and 7 m potassium hydroxide. Decomposition was followed by measuring the oxygen pressure above the ferrate solutions. He found the decomposition rate was not effected by light and only small changes in the decomposition rate was found for solutions containing glass beads or a Fe (III) solid phase. As previously reported, Magee found decomposi- tion of ferrate in pH 9.2 and 9.6 solution at 250 second- order in ferrate and this author found comparable rate constants at 25°. Magee used 1 x 10.3 M ferrate 3 M ferrate solutions and for pH 9.2, l - 1.5 x 10- solutions, found a large variation in the rate constant, a higher rate being obtained with a higher ferrate concentration similar to that found here for the pH 9.1 solution at 40°. Magee also measured the decomposition rate in pH 13.8, 4.9 and 7.2 sodium hydroxide solutions, but 163 did not attempt to control the ionic strength. He suggested that the decrease in the decomposition rate with the higher sodium hydroxide content was due to the higher ionic strength of the solutions and that the decomposition was first-order in ferrate and first or zero-order in hydroxide ion. The present work suggests that the rate decreases with higher ionic strength and hydroxide ion concentration. Comparison of Magee's decomposition rates at 250 (ionic strength 0.15) with rates found here at 400 (ionic strength two) show his values are about one hundredfold lower for pH 10.6 and 11.1. For 7.2 m sodium hydroxide, Magee reported a rate constant of -1 4 x 10'4 min for 25°, while in the present work, a value of 2.0 x 10-3 min- was found in 7 M sodium hydroxide at 400 (10‘3 M ferrate). However, Magee found a much lower rate in 4.9m sodium hydroxide (kexp. = 7 x 10'5 min-1) at 25° than in the 7.2 m solution. In the present work, a general decrease in the decomposition rate was found as the sodium hydroxide content increased. Wronska (20) studied the decomposition of 3 - 1 x 10"2 M ferrate oxidimetrically in 2.5 x 10- initially neutral solutions at 20° and 30°, without controlling the pH or ionic strength. She suggested 164 the data fit the rate equation, - k(a-x)2 _.__§____ rate where (a-x) ferrate (VI) (OH'). and x However, in addition to the ionic strength change, the pH of the solution changed, and for the higher ferrate (VI) concentrations, the solution was basic before her measurements were started. Since a change in mechanism occurs between pH 9 - 10, it is doubtful if meaningful results can be obtained in this manner. Magee (22) attempted to correlate Wronska's data with his, but was not successful and critically discussed her work. Wronska and Trzebiatowska (19) studied the decomposition of ferrate in 7 - 10 M potassium hydroxide. Ferrate concentrations of 6 x 10’4 to 2 x 10-3 M were studied at 200 and 30°. They found the decomposition to be first—order in ferrate and stated that a deviation from first-order occurred with the precipitation of Fe (III). Rate constants reported were 50 to 100 times faster than found by Magee (22) at 25° in 7 m potassium hydroxide. For 7 M potassium hydroxide, these authors reported a rate constant of 6.0 x 10‘2 min"1 at 20° and 0.10 min-1 at 309. In 165 this present study, for 10-3 M ferrate solutions in 7 M sodium hydroxide, a rate constant of 2.0 x 10"3 was found at 40°. For 1 x 10.4 M ferrate solution at 40°, a rate constant of 0.13 min.1 was found here, but curves were obtained for higher ferrate concentrations up until precipitation of Fe (III) occurred. Magee (22) found that the decompositiOn rate of ferrate was similar in 7 M sodium hydroxide and 7 M potassium hydroxide solutions. Thus, one explanation for the faster rates reported by Wronska and Trzebiatowska would be that the initial slopes of the plots were used for determining the rate constants. However, the rate constant was reported to be constant for 90% of the decomposition. Insufficient data.did not allow a more definite conclusion regarding the rate values reported by Wronska and Trzebiatowska. Wronska and Trzebiatowska suggested a mechanism for the decomposition of ferrate in 7 — 10 M potassium hydroxide. They proposed that a ferrate ion coordinates with two hydroxide ions, which decomposes to form a Fe (IV) species plus an oxygen atom. The authors then suggested that the Fe (IV) may decompose by some hydroxyl radical mechanism. 11L? E veg..." GI. _ 8153‘ SUMMARY AND CONCLUS ION 167 Methods g§_Analysis The decomposition of potassium ferrate (VI) in aqueous solutions may be followed by several different methods of analysis. The oxidizing power of ferrate (VI) may be used to determine the concentration of ferrate (VI) oxidimetrically. Measurement of the oxygen formed in the decomposition may also be employed as a method of analysis. Under certain conditions, the absorbance of the colored ferrate (VI) ion is a satisfactory method of following the decomposition. The modified chromite method appeared to be the most versatile and reliable method of analysis. The method was not effected by the presence or formation of an Fe (III) solid phase and was adaptable to the entire concentration range studied. Although the original titrimetric chromite method (26,27) is expected to give more accurate values, it requires larger amounts of ferrate, is more time consuming and is not adaptable for dilute (10'5 - 10‘4 M) ferrate solutions. Also, when the data are graphically represented, the increased accuracy is not required. The disadvantages of this method of analysis was the labor involved in the analysis, which generally resulted in a smaller number of experimental values being obtained. 168 The direct photometric method was restricted to homogeneous solutions and to low initial ferrate concentrations. Formation of a turbid solution prevented the use of the method for several of the solutions used in the study and restricted the ferrate concentration that could be studied in many of the other solutions. A second disadvantage was that constant stirring was not possible. However, the method had the advantages that the ferrate (VI) absorbance is continuously recorded and that only periodic inspection is required. The oxygen evolution method followed the decomposition by periodically measuring the volume of oxygen evolved from a ferrate solution. The method could also be adapted to follow the pressure change above a ferrate solution, but would require correcting for the increased solubility of oxygen as the pressure increased (22). The method permitted collection of a large amount of data for each run and required little attention once the system was established. In addition, it was necessary to presaturate the solutions used in the decomposition with oxygen and stirring was required throughout the decomposition. For many of the solutions, a low decomposition rate was obtained using this method and the method 169 appeared to be limited to measuring slower decomposition rates. When used for measuring decomposition rates, it is advisable that the results be compared with those from another method. Fggpgte Decomposition The decomposition of potassium ferrate (VI) in aqueous solutions is complex and subject to several experimental variables. Oxygen, Fe (III) and hydroxide ion formed as decomposition products and Fe (III) and hydroxide ion are known to effect the rate of decomposition. The decomposition rate is also effected by the ionic strength of the solution, and it was necessary to control this variable. Two sets of solutions were used in this study, the first set consisted of buffer and sodium hydroxide solutions from pH 7.7 to 2 M sodium hydroxide, adjusted to an ionic strength of two. The second set consisted of l - 7 M sodium hydroxide solutiors in which an attempt was made to adjust the ionic strength to seven with sodium perchlorate. V For the pH 7.7 - 2 M sodium hydroxide solutions, the data indicated that a lower decomposition rate was found with a higher ionic strength. For the 170 l - 7 M sodium hydroxide solutions, 5 x 10-5 - l x 10.4 M ferrate solutions were not effected by the sodium perchlorate content of the solution. For 10"3 - 10.2 M ferrate solutions, a lower decomposition rate was found for solutions with a high ionic strength and was the result of an inhibiting effect of Fe (III). Decompositions in the l - 7 M sodium hydroxide solutions which contain sodium perchlorate, more closely resemble the decomposition in 7 M sodium hydroxide than do sodium hydroxide solutions that do not contain sodium perchlorate. The decomposition rate of potassium ferrate (VI) was also subject to the alkalinity of the solution. Above pH 12, the rate of decomposition decreased as the hydroxide ion concentration increased. Decomposi- tion rates in pH 13.8 to 2 M sodium hydroxide solutions (ionic strength two) indicated that the decomposition was inhibited by (OH)7. For the l - 7 M sodium hydroxide solutions, maintained at a higher ionic strength, the same inhibition was found for dilute ferrate (VI) solutions (10"5 - 10"4 M). With the 10‘3 - 10.2 M ferrate solutions, the decrease in rate with hydroxide ion was much greater, the rate appearing to be inhibited by (OH)2. The difference in results for the two ferrate concentrations was attributed to 171 3 - 10"2 M the inhibiting effect of Fe (III) in the 10' ferrate solutions. For decompositions in solutions above pH 13, the decrease in decomposition rate is interpreted as reflecting the increased stability of ferrate (VI) and possible intermediates, rather than involving protonated species. Even though the kinetic rates cannot be derived from reduction potentials, the reduction potential of ferrate (VI) does indicate an increased stability at higher alkalinities as shown by the following: FeOZ + 3é-+ 2H20 = FeOE + 40H7 E0 = 0.72 volts (12). The above equation also indicates that the oxidation of Fe (III) occurs more readily in concentrated hydroxide solutions. The inhibiting effect of Fe (III) is believed to result from oxidation to a higher oxidation state during the decomposition. From pH 7.7 to pH 12, the decomposition rate shows two different trends. From pH 7.7 to 9.5, the decomposition rate decreased as the pH increased. 5 For both dilute ferrate solutions (10' - 1074 M) and 10.3 M ferrate solutions, the decomposition was first-order in hydrogen ion. For the 10'3 M ferrate solutions, the decomposition was also second-order in 172 ferrate. The pH 7.7 - 9.5 solutions, which were 10-3 M in ferrate, were the only solutions where the decomposition was second-order in ferrate. Decomposition of dilute ferrate solutions was first-order in ferrate for pH 7.7 - 9.5 solutions. From pH 9.5 to pH 12, the decomposition rate increased rapidly, reaching a maximum in the vicinity of pH 11.5. The data indicated that from pH 7.7 to 9.5, the ferrate species involved in the decomposition was a mono-protonated species, while above pH 9.5 only the dinegative ferrate ion was present. No spectro- photometric evidence in the visible spectral region was found to support this conclusion. The ultra- violet region could not be investigated due to the absorption of the buffers in this region. Since the decomposition rate for 10"3 M ferrate solutions was second-order in ferrate between pH 7.7 - 9.5, it is expected that an interaction between a protonated species and the dinegative ion would occur more readily than an interaction between two dinegative ions. The decomposition rate in the vicinity of pH 9.5 is very low, being lower than the rate in 2 M sodium hydroxide. The low decomposition rate very likely resulted from a change in the mechanism at this pH value. The increase in the decomposition rate from 173 pH 10 to 12 is quite rapid and there does not appear to be a constant dependence on pH for this range. After pH 12, the decrease in rate with alkalinity is uniform. The effect of Fe (III) on the decomposition rate of ferrate (VI) can be generalized as occurring in solutions with and without a solid Fe (III) phase. To maintain a homogeneous solution, it was necessary to limit the ferrate (VI) concentration to less than 3 x 10"3 M ferrate for many of the solutions. With some of the solutions, the ferrate (VI) concentration was limited to l x 10"4 M. In the solutions where an Fe (III) precipitate did not form, the effect of Fe (III) on the decomposition rate could be determined. For pH 10 to 2 M sodium hydroxide solutions, the effect of Fe (III) on the decomposition rate depended on the initial ferrate (VI) concentration and the total amount of Fe (III) present in solution. For very low ferrate (VI) concentrations (5 x 10"5 M), the decomposition rate was first-order in both ferrate (VI) and in Fe (III). From 1 - 3 x 10"4 M ferrate (VI), I the decomposition was first-order in ferrate and one- half-order in Fe (III). Higher initial concentrations of ferrate (VI) appeared to be first-order in ferrate. There were indications of a lower decomposition rate 174 as the Fe (III) approached its solubility limit. The ferrate (VI) concentration for which the above Fe (III) effects were found, varied slightly with the alkalinity of the solution. As the hydroxide concentration increased, the decomposition rate appeared to indicate a decreasing dependence on Fe (III). For 1 - 7 M sodium hydroxide solutions (high ionic strength), an acceleration by Fe (III) was not found although there was some indication that a very rapid decomposition took place when ferrate (VI) was added to a solution containing Fe (III). For the l - 7 M solutions, Fe (III) was found to inhibit the decomposition rate. For solutions where a rapid decomposition occurred upon addition of ferrate (VI) to a solution containing Fe (III), the remaining ferrate solution decomposed at a much lower rate than expected. The decomposition rate was found to be first- order in ferrate (VI) and inhibited by Fe (III) for initial ferrate concentrations greater than 1 x 10.4 M or solutions that initially contained Fe (III). For initial ferrate concentrations less than 1 x 10-4 M, the decomposition was first-order in ferrate (VI). With solutions that did not contain sodium perchlorate, the inhibiting effect of Fe (III) was not observed until approximately 4 M sodium hydroxide. Decomposition 175 of ferrate (VI) in solutions with a lower sodium hydroxide content than 4 M, resembled decompositions in l - 2 M sodium hydroxide solutions (with ionic strength of two). For initial ferrate concentrations less than 1 x 10'4 M which did not contain Fe (III) initially, the decomposition was first-order in ferrate and approximately the same decomposition rate was found for sodium hydroxide solutions with and without sodium perchlorate. The varying dependence of the decomposition on Fe (III) indicates that several reaction steps were involved in the decomposition which may be reversible and depend on the concentration of the different species and the alkalinity of the solution. The Fe (II) - Fe (III) peroxide system serves as an example of such a system. George and Hargrove (43) reported that the dependence of the rate on the peroxide concentration changed with the peroxide/iron ratio. As the peroxide concentration increased, the dependence on peroxide changed from second-order,to less than second-order,to first-order and finally to zero. Decomposition of 10"3 - 10"2 M ferrate (VI) solutions resulted in a solid Fe (III) phase forming in the solution. For decomposition in heterogeneous systems, two separate cases were found. For most 176 of the solutions, the presence of additional Fe (III) solid did not appear to effect the decomposition rate. For these solutions, once an Fe (III) phase formed, the Fe (III) concentration in solution very likely remained constant during the decomposition. For these solutions, the decomposition was first-order in ferrate (VI), except for pH 7.7 - 9.0 range where the decomposition was found to be second-order in ferrate. For the pH 9 - 10 range (ionic strength two) and sodium hydroxide solutions above 3 M (ionic strength seven), addition of an Fe (III) solid phase increased the decomposition rate of the ferrate (VI). The heterogeneous mechanism was not thoroughly investigated but it appears to involve Fe (III), as glass beads did not alter the decomposition rate to the same extent as the Fe (III) solid. It is of interest to note that the heterogeneous reaction was noted for solutions with a slow decomposition rate. It is possible that in many of the solutions, the decomposition rate in solution is more rapid than the heterogeneous rate. For pH 9 - 10 range, it appeared that the decomposition was second-order in ferrate, but appeared first-order in ferrate when an Fe (III) solid phase accumulated. Addition of solid Fe (III) phase resulted in the decomposition appearing first-order in ferrate, and the 177 rate was dependent on the amount of Fe (III) solid present. With moderate amounts of Fe (III) solid, the rate increased 20 to 30 times over that found starting with only ferrate (VI). It is believed that the low decomposition rate resulted from a change in the decomposition mechanism between pH 9 - 10 and that the Fe (III) solid provided a more accessible decomposition path. For the concentrated sodium hydroxide solutions, the rate was approximately doubled when additional Fe (III) solid was added. The decomposition with and without additional Fe (III) solid appeared first-order in ferrate. In considering the mechanism of ferrate (VI) decomposition, other authors (14,19) have suggested that hydroxyl radicals were involved in the decomposi- tion of ferrate (VI) in concentrated hydroxide solutions. The formation of radicals from a strong oxidant in a hydroxide solution seemed reasonable. However, no evidence for hydroxyl radicals was found. Gump (3), Magee (22) and this author found no acceleration of the decomposition with light of the visible spectrum. Only slight changes in rate were found when glass beads or glass wool were placed in the ferrate (VI) solutions and for many solutions, no change in rate was found with larger amounts 178 of Fe (III) solid phase in the ferrate (VI) solutions. Also, oxygen or nitrogen did not appear to change the decomposition rate, except in the oxygen evolution method where a physical effect was found. A large change in rate under these various conditions was expected if a free radical mechanism was operative (44). A single electronparamagnetic resonance spectrum was run on a ferrate (VI) solution in 7 M sodium hydroxide and again no evidence for radicals was found. It has been reported (45) that hydroxyl radicals formed in an alkaline solution react with benzene to form biphenyl or phenol. This test was used with 7 M sodium hydroxide solution, 10"2 M in ferrate,which was rapidly mixed for several hours with 30 ml of benzene. The resulting benzene solution was concentrated and this concentrate was analyzed by gas chromatography. No evidence for any products except water or benzene was found. However, it was possible that the above method of analysis was not sensitive enough to detect trace amounts of these materials. Although radicals may be involved in the decomposition of ferrate (VI), the lack Of evidence for a radical mechanism resulted in a different approach being taken to explain the decomposition. 3 The decomposition of 10- M ferrate (VI) solutions, 179 pH 7.7 - 9.5,are second-order in ferrate and first— order in hydrogen ion. Magee (22) studied pH 9.2 and 9.6 solutions at 25° and suggested the following mechanism, _ k l. Fe042 + H+.—_1-‘ HFe04' k-l 2 HFeo’I- FeO 2" 3.2. HFe(VI)0 0' + FeO 2‘ ° 4 4 4 3 _ - 2- k 3. HFe(VI)040 + 0H 3, FeO3 t 02 t H20 2 2- k _ - 3- 4. F803 + FeO4 .__41 F602 + Fe(V)040 3- _ 5. Fe(V)040 _k5_. Fe033 + 02 3_ _ _ o = 2 6 Fe03 + H20 FeO2 -+ 0H 7. Fe02- I- 2320 : Fe(OH)3 + 03" The second step was considered the rate step while other steps were fast. Products of reaction 2 and 4 are peroxoferrates. Accounting for the disappearance of ferrate (VI) and applying a steady state approximation gave the following, 2... —d o - + ' (F2: ) - k1 (Fe042_) (H ) - L1 (”904 ) + k2(HFe04‘) (Fe042 ') + k4(Fe03 ") (Fe042) A . (HFe(VI)04OT'= k2(HFe0;) (Fe042-) k3 (OH') (FeO 2”I = k2(HFe04") (8e042') + k3(HFe(VI)040-) (011‘) 180 Substitution gives: -d(FeQ42') : 3k2k1 (FeO 2")2 (H+) dt '3:““ 4 -1 - 2 The above mechanism appears to desribe the decomposi— tion of 10'3 M ferrate (VI) in the pH 7.7 - 9.5 range and the derived rate equation is identical with the equation obtained experimentally. The peroxoferrate species proposed in the mechanism seem reasonable, considering the oxidizing properties of ferrate and the pH of the solutions. For dilute ferrate solutions in pH 10 to 2 M sodium hydroxide, the following mechanism appears to desribe the decomposition. l. FeOZ + Feoz‘ kJ rec; + Fe03 = : k - 3- 2. Fe04 + Fe03_2..Fe02 + Fe(V)040 3- 3- 3. Fe(V)040 k3 FeO3 + 02 _ 3_ z 4. F603 + F803 R4 2 £603 3.. 5. Fe0 Feoz’ + 20H- Fe(OH)3 + OH- The Fe (III) species actually involved in the decompo— sition very likely varied over the hydroxide ion concentrations that were studied, but for simplicity are represented here as FeOZ‘ and Fe033' in step 1 and step 4 respectively. In the more concentrated hydroxide solutions,the species very likely was Fe02" From the above steps, 1;..- 181 2- 2- _ :Qi£224__l = k1(Feo ) FeO ) + dt 4 Q 2 k = = 2(Fe04) (Fe03), By steady state approximation, and substitution (F603-) 2 k1 (FeO 2') E2 4 3- (Fe032-, : 3k1(Fe03 ) k2 2 _ k3 -d(Feo42-) - 4k (F o 2") dt' _ 1 e 4 (Fe III) The above equation is the same form as the experimental equation used for several of the dilute ferrate solutions. Also used for the experimental plots was an equation first-order in ferrate (VI) and one- half-order in Fe (III). However, this equation fit the data up until approximately 3 x 10"4 M ferrate (VI) after which the decomposition appeared to be only first-order in ferrate (VI). In the mechanism given for dilute ferrate (VI) solutions, if an additional reaction step is considered, a second equation can be obtained. After step two, where a peroxoferrate (V) species is shown, if the intermediate reacts with Fe (III), then 2'. Fe(V)0403- + Fe02‘ k2' Fe043- + FeO3 7. Feo43‘ + H20 = Feo3‘ + 20H- filmy-ko- ri..lll. _ .i will": 'l.ii 182 and ~ ‘ _ , _ 3k1(FeOZ) [k2 (F602) + k3] k2 Lk3 - 3k2° (FeOEfl _ _ 2_ _ o FeOS = 2- _ dt ' - 3k ' FeO‘ k3 2 2 The above equation shows that as the Fe (III) concentra- tion builds up in the solution, the decomposition rate will depend on the values of k1, kz', k3 and the Fe (III) concentration. The value of 3k2'(Fedé) is expected to remain smaller than k3. It was found experimentally that in the pH 10 to 10.6 range, as the initial ferrate (VI) concentration was increased the rate became less dependent on the Fe (III) concentration and eventually a dependence on Fe (III) was not evident. A similar mechanism may be operative in pH 7.7 to 9.5 solutions for 10"5 - 10’4 M ferrate (VI), as variation of the decomposition rate with the initial ferrate concentration was also found for these solutions. However, the Fe (III) solubility was too low to permit an extensive investigation in this pH range. fling- .1 U 55.ni 183 For the concentrated sodium hydroxide solutions, Fe (III) was found to inhibit the decomposition. This suggested that Fe (III) was oxidized to a higher oxidation state during the decomposition. Also, in the high hydroxide concentration, it seemed reasonable that a reaction might take place between ferrate (VI) and hydroxide ion. The following mechanism was developed to explain the decomposition in concentrated hydroxide solutions. _ - _ + 1. Fe042 + OH kl Fe(Iv)o4o4 + H 2. Fe(IV)04O4- + FeOz’ k2 Fe032- + Fe043' 4- 4- 3. Fe(IV)O4O k3 Fe03 + 02 4. F6042- + Fe034‘ k4 Feo44' + Feo32' 5. Fe042‘ + Fe032“.35AFe(V)0403' + FeOZ' 6. Fe(v)o4o3- + Fe02"‘E§lFe043' + Fe03’ 7. F8(V) 0403- 3... 8. Fe043' + Fe02‘ k8 2Feo§ Applying a steady state approximation, collecting terms and substitution, gave the following: 4- - (Fe03 ) - k3 k1(OH) —] k4 k2(Fe02‘) + k3 k2