V O L U M E 26, NO. 6, J U N E 1 9 5 4
1025
by reversing the gas flow, which allows for shorter Tygon connections with the rePt, of the apparatus because the gas control valves on the instrument are bypassed. -4modification of the procedure is to use a detachable cold trap and expand the gases, after removal of the oxygen, into a m:tss spectrometer ( 1 ) . Because of the high-ronibustion temperature attainable, no accelerator or flux need be used in most cases for millings, turnings, or drillings. Titanium and zirconium are exceptions. Complete combustion in some cases requires more time than is available if the oxygrn flow is maintained at, the prescribed flow rate of about lo00 nil. per minute because of the size of the gas buret. h simple modifiestion in the procedure corrects this difficulty and indicates t,he flexibility of the equipment. After t,he start of the ignition, the oxygen flow is regulated at the pinch clamp to supply just sufficient oxygen t o support t,he conihustion. This is accomplished by JTatching the liquid level in the gasometer and maintaining it stationary for the desired dur:ition of t.he combustion prriod. If the oxygen flow rate is enough, the leveling solution will start to rise in the gasometclr which wrvee as a very convenient flowmeter. By this means, combustion times of as long as 5 minute8 werc used to effect the complete oxidation of tin, copper, lead, and molybdenum. Without. this precaution, the materials are incompletely oxidized and the carbon values are low and erratic. TVlien the metals are completely converted t o the corresponding oxides, there is little or no inductive coupling and rapid decrease in the temperature of the sample, as shown by its color, indicatty the end of the combustion. The combustion products are then swept into the gasometer and the determination is continued. The very reactive metals, such as titanium and zirconium, represent the opposite extreme. These metals flash-burn too quickly when burned alone and produce very erratic results. They are hwt handled by introducing a flux which will maintain
the combustion for about 5 minutes to ensure the complete oxidation of the carbides. This flux should alloy with the sample t o slow down the oxidation and form a homogeneous melt. . After some experimentation manganese was found to be the ideal bedding material for this purpose. B y itself, it burns more smoothly and at a higher temperature than iron, but more important, it forms homogeneous melts with titanium and zirconium. Also it is one of the lowest in carbon content, so that the correction required is small. Because only a small quantity of crystal bar titanium was available, 300-mg. samples were used with 0.5 gram of electrolytic manganese as the flux. S o zirconium values are reported because the available material was so high in carbon that it wac outside the range covered in this paper. Nevertheless, good controllable burnings were obtained with zirconium (and titanium) on 1-gram samples plus 1 gram of manganese. LITERATURE CITED
(4) (5)
Hickam. \T. \I., .kNAL. CHEXI., 24, 362 (1952). Horton. \Ir.S., and Brady, J., Ibid.. 25, 1891 (1953). JIurray, W.AI., and dshley, S. E. Q., IXD.ENG.CHEM...ixt1.. ED., 16, 242 (1944). lIurrag, W.lI.,and Siedrach, L. IT., Ibid., 16, 634 (1944). Pepkowita, L. P., and Chebiniak, P., ANI.. CHEM.,24, 889 (1952). ~ - , Stanley. J. K., and Yensen, I. D.. IND.ENG.CHmf., . i s < r , . ED.. 17, 699 (1946). Wells, J. E., .J. Iron Steel Inst., 166, 113 (1950). Wooten, L. A , . and Guldner, W.G., IND.ENG.CHEY..-1st~. ED.,14, 836 (1942). Yensen, T. D., Trans. Electrochem. S’oc., 37, 227 (1920). Ziegler, S . -1.. Ihid., 56, 231 (1929). \ -
(6)
(7) (8)
(9) (10)
RECEIVED for review January 7 , 1954. Accepted April 7 , 1954.
The Knolls Atomic Power Laboratory is operated b y General Electric Co. for t h e .Itornic EnerEy Commission. This work x a s carried out under Contract \1--31-109 Enq-jZ.
Quadrivalent Uranium as a Reducing Titrant R O N A L D BELCHER, DEREK GIBBONS, and T H O M A S S. WEST Department o f Chemistry, University
o f Birmingham,
Birmingham 75, England
Quadriialent uranium has a redox potential of -0.334 volt (10) in O . l \ - acid solution and is, therefore, a moderate reducing agent. It is fairly stable to atmospheric oxidation. Quadrivalent uranium, as a reducing titrant, is one of the most stable titrants that can be used for the direct titration of ferric iron. Although i t can be used for seieral other titrations, i t has no advantages over existing reagents, and, in most cases, is less conbenient.
S
IIVERAI, reagents utilizing a less common valency state of the element concerned have found wide applications in titrimetric analysis. Chroniium(I1) ( 5 ) , titanium(II1) ( 7 ) , and vanadium(II1) (1.5) salts have long been used as reducing titrant,s and, more recently, tungsten(II1) (lb), copper(II1) (I), nirkel(II1) ( I S ) , silver(II1) ( X I ) ,manganese(II1) ( S ) , and molyhdenum(V) ( 1 6 )have been examined. Quadrivalent uranium i p one of the few elements, in an unU S U ~valency stat>e,which has not been examined extensively as n titrimetric reagent, although there is considerable information in t,he literatuie (8, f / i ) ronrerning the titration of uranium(IV) by other reagents. The only work descrihed hitherto, in which quadrivalent uranium is used ai; n titrant, is by Vortmann and Binder ( 1 7 ) ,
n ho suggested the use of uranous sulfate as a reagent for the determination of nitrate, chlorate, chromate, manganese dioxide, 4 visual method was used for the detection of and ferric iron. . the end point (the disappearance of a ferric thiocyanate color), and consequently only ferric iron could be determined directly. The other oxidants \!ere determined in one of two ways: An excess of ferrous sulfate solution was added, and the ferric iron formed was titrated with uranous sulfate 4 n ewess of uranous sulfate solution was added, and the ewess m s back-titrated with ferric iron. The present authors have extended the work of T-ortmann and Binder and have examined the possibility of titrating several reducible substances. The end points in these titrations were detected potentiometrically, using a Mullard “magic-eye” potentiometric bridge (llullard Ltd., Century House, Shaftesburjdve., London W.C.P., England). illthough quadrivalent uranium can be used for several titrations, it has little advantage over esisting reagents: in most cases, it is less convenient. EXPERIMENTAL
Preparation of the Reagent. The uranous solution was pre-
pared under the conditions described by Willard and Diehl About 43 grams of analytical reagent grade uranyl acetate were dissolved in 500 ml. of 4 5 hydrochloric acid and the solution was allowed to ( 1 9 ) for the quantitative reduction of uranyl salts.
ANALYTICAL CHEMISTRY
1026 flow down a silver reductor (18)a t about 5 ml. per minute. The reduced solution was diluted to 2 liters to give an approximately 0.1N uranous solution in LV hydrochloric acid. Uranyl acetate was used as starting material for the preparation of the reagent, rather than the chloride, because the chloride was not commercially available in a sufficiently pure state. Uranyl nitrate could not be used, because the nitrate ion interferes in the titration of quadrivalent uranium with dichromate (14). Stability of the Reagent. Vortmann and Binder (17 ) prepared their uranous sulfate solution by reduction of uranyl sulfate with zinc in sulfuric acid. The solution was stored in a stoppered dark bottle and found to be 2570 oxidized after 28 days. The present authors also stored the uranous solution in a dark bottle and fed it into the bottom of a buret by means of a siphon and two-way tap system. The solution was, therefore, open to the air. The reagent was titrated against standard potassium dichromate solution, each morning over a period of 10 days. No appreciable oxidation took place over the first 24 hours, but the titer increased steadily thereafter.
oxidation of the hydrochloric acid by the permanganate a t the high temperature, with concomitant loss of chlorine. This still occurred when the hydrochloric acid medium n as replaced by sulfuric acid, because there was still hydrochloric acid in the uranous solution used as titrant. When the titration was carried out in the cold, no change in the oxidation potential of the system occurred, but the permanganate was decolorized. The color change a t the end point was from red to greenish yellow through yellow-brown. This gave satisfactory results if the end point was taken when the first tinge of green appeared in the solution, However, when the titration was carried out hot in the presence of 20 ml. of Zimmerman-Reinhardt preventive solution, good r e sults were obtained. The titration of ferric iron was more troublesome. Even a t 60" C. the reaction was somewhat sluggish, causing indistinct end points and slightly high results. It was found preferable to add the titrant slowly near the equivalence point (0.1 ml. per minute), rather than to work a t higher temperatureq as in the titration of quadrivalent uranium with ferric iron (14). Ferricyanides were titrated in a way similar to ferric iron, but the reaction was even more sluggish, presumably owing to the stabilizing effect of the cyanide ions. However, provided that the titration was carried out slowly near the equivalence point, it was possible to obtain satisfactory results. TITRATION OF OTHER REDUCIBLE SUBSTANCES
PH
Figure 1. Variation of Reducing Power of the Uranous Reagent with pH
A reductor buret ( 4 ) using silver or lead as reductant, was not found suitable, under ordinary conditions, for maintaining the uranium solution in the quadrivalent state. When the reagent was stored under a layer of petroleum ether (boiling point 100" to 120" C.), no oxidation took place even after several weeks. A layer of,petroleum ether was also placed on top of the solution in the buret, to minimize oxidation, but this was not essential, provided that the buret was filled with fresh s o h tion for each titration. The layer of petroleum ether, however, gave a sharp definition of the meniscus in the buret. Standardization of the Reagent. The uranous solution was standardized by titration against a standard solution of potassium dichromate, prepared by weight from the oven-dried analytical reagent grade salt. Twenty milliliters of 0.1-T potassium dichromate solution were run into a 400-ml. beaker from a buret, and 10 ml. of 2N hydrochloric acid were added. The solution was stirred magnetically and titrated potentiometrically with the uranous solution, using a bright platinum indicator electrode and a saturated calomel reference electrode. The reaction was found to be somewhat sluggish a t room temperature, making the end point rather indistinct, but when the titration was carried out a t 60' C., satisfactory and consistent results were obtained. Similar results were also obtained using 0.02-V solutions. TITRATION OF VARIOUS IONS
Cerate, permanganate] and vanadate solutions were determined in a way similar to the titration of dichromate. It was again necessary to work a t elevated temperatures, but 60" C. was found to be sufficient, provided the titration was performed carefully in the vicinity of the end point. I n the titration of permanganate, however, low results were obtained owing to
Chlorate, persulfate, and peroxide were reduced slowly by the uranous solution even a t elevated temperatures, and satisfactory end points could not be obtained. In hot solution oxidation of the hydrochloric acid occurred, which caused low results, All three radicals could, however, be determined, over a range of concentration, through ferrous ion. A suitable excess of ferrous ammonium sulfate solution was added to a known amount of the oxidant and the mixture warmed to 60" C. for 5 to 10 minutes. The solution was made 1 S with respect to hydrochloric acid, and then titrated with the uranous solution, as for ferric iron. Consistent results were obtained, and the amounts of uranous solution consumed corresponded to the calculated values. Vortmann and Binder ( 1 7 ) determined nitrate in this way by boiling the nitrate solution with ferrous sulfate in the presence of hydrochloric acid in an atmosphere of carbon dioxide. An attempt \vas made to apply the more convenient conditions employed by Leithe (11), but the high acidity made the subsequent titration with uranous solution too sluggish. When the acidity was reduced to approximately lNl by the addition of sodium bicarbonate, sodium carbonate, or sodium hydroxide, a satisfactory end point was still not obtainable] because the high electrolyte concentration upset the redox potentials of the system. Finally, as the potential ranges involved in the reduction of ferric ion and dichromate were sufficiently weil separated, it was possible to determine these two ions in the presence of each other. VARIATION OF REDOX POTENTIAL WITH ACIDITY
Searching studies of the redox potentials of the uranium(1V)uranium(V1) system have already been published (6, 9), but for the sake of completeness, the redox potential of the system was determined over a range of pH. -4 suitable amount of potassium dichromate solution was titrated with uranous chloride in the usual way, and then an equal amount of uranous solution was added in excess. The acidity was adjusted to the required pH, and the potential of the system was measured using bright platinum and saturated calomel electrodes. The potential with respect to the 'hydrogen electrode was then calculated, This was repeated over a range of pH. The graph obtained (Figure 1 ) shows a fall in potential (and therefore an increase in reducing power) with increasing pH. It
V O L U M E 26, NO. 6, J U N E 1 9 5 4
1027
would appear that a t an even higher p H the uranous solution would be a still more powerful reductant, but as p H 2 was approached the solution became dark brown in color, and beyond this, precipitation of the uranium occurred. CONCLUSIONS
Quadrivalent uranium is only a moderate reducing agent in acid solution, but can be used for the direct titration of thestronger oxidants and ferric iron. Weak oxidants may be determined through their oxidation of ferrous iron. Ferric iron may also be determined in the presence of dichromate. Quadrivalent uranium is the most stable reagent that has been used for the direct titration of ferric iron, apart from mercurous nitrate ( 2 ) , but the method is less convenient and holds no advantage over those involving existing reagents. ACKNOWLEDGMENTS
The authors are endebted to the Department of Scientific and Industrial Research for grants enabling this work to be undertaken. LITERATURE CITED
(1) Beck, G., Mikrochemie ver Mikrochim. Acta, 35, 169 (1950). (2) Belcher, R.. and West, T. S.. A n a l . Chim. Acta, 5 , 260, 268,
360 (1951).
(3) Ibid., 6, 322 (1951). (4) E’laschka, H., Ibid., 4, 242 (1950). ( 5 ) Flatt, R., and Sommer, F., Helv. Chim. Acta, 25, 654 (1942). (6) Heal, G., Trans. Faraday SOC.,45, 1 (1949).
(7) Knecht, E., and Hibbert, E., “Kew Reduction Methods of Volumetric Analysis,” London, Longmans, Green & Co., 1925. (5) Kolthoff, I. M.,and Lingane, J. J., J . Am. Chem. SOC.,55, 1571 (1933). (9) Kraus, K. A , and Nelson, F., Ibid., 71, 2517 (1949). (10) Latimer, W. lI.,“Oxidation Potentials,” p. 304, New York, Prentice-Hall, Inc., 1952. (11) Leithe, W., ANAL.CHEM.,20, 1052 (1945). (12) PFibil, R., and Usel, R., Collectton Crechoslov. Chem. Commune., 10, 330 (1935). (13) Rly, P., and Sarma, B., .Vatwe, 157, 627 (1946). (14) Rodden, C. J., “Analytical Chemistry of the Manhattan Project,’’ p. 70, New York, McGraw-Hill Book Co., 1950. (15) Russell, -4.S., J . Chem. SOC.,129, 497 (1926). (16) Tourky, A. R., Farah, 31.Y., and Shamy, H. K. El, Analyst, 73, 255, 262, 266 (1945). (17) Vortmann, G., and Binder, F., Z . anal. Chem., 67, 269 (1925) (18) Walden, G. H., Hammett, L. P., and Edmonds, S. hI., J . Am. Chem. Soc., 56, 350 (1934). (19) Willard, H. H., and Diehl, H., “ddvanced Quantitative dnalysis,” p. 98. Xew York, D. Van Nostrand Co., 1944. (20) Yost, D. >I., J . Am. CRem. Soc., 48, 152 (1926). R E C E I ~ Efor D reiiew July 7, 1953. Accepted March 6 , 1954.
Infrared Absorptiometry for Quantitative Determination of Boron Hydrides in Presence of Pentaborane SMITH,and ROBERT S. M C D O N A L D Schenectady, N. Y.
LEWIS V. MCCARTY, GEORGE C. General E/ectric Research Laboratory,
Physical methods of determination for the boron hydrides are much easier to apply than conventional chemical methods. Of the available physical methods infrared absorption in the gas phase offers a number of advantages. Data are presented for the determination of the four component system containing diborane, tetraborane and dihydropentaborane in a relatively larger amount of pentaborane. The data are combined into equations which permit the direct calculation of the pressure in m m . for each boron hydride from measurements of the absorbance, D, at four wave lengths.
- 3.49 Dj.51 + 14.81 D6.16 68.55 D1.65 - 1.45 Dj.54 + 0.47 D6.16
+ 0.52
PB,H,= - 0.58 D4.o;
PB,H,,= - 3.64 D4,0jf
04.65
+ 14.65 D4.65 + 164.10 Dj.54 + 0.45 D6.15 P B ~ H= , , 39.86 D1.06 - 74.85 + 2.48 Dj.54 - 5.12 D6.15 PB,H,= - 7.25 Di,dj
04.65
At a total pressure of 100 m m . the minimum detectable amount of diborane isabout0.2mole yo; for tetrahorane i t is about 0.6 mole yo; and for dihydropentaborane i t is about 0.7 mole %.
T
HE quantitative and qualitative chemical determination of
boron hydrides is a very difficult matter because of the similarity of the chemical properties of the compounds. With further research it may be possible to determine an individual material chemically, but at present physical methods of analysis are much easier to apply. Four techniques have been used in this laboratorv and certain advantages can be claimed for each. To set
infrared ahsorption analysis in its proper perspective with respect to these other methods, a brief review of each method is presented here. Cryoscopic determination of pentaborane content is an excellent method for checking the quality of purified pentaborane and can probably be used as an analytical tool for material which contains a- much as 20 mole % impurity. However, the method tells nothing of the kind of impurity that is present. The mass spectrometer is an excellent method for qualitative identification of boron hydrides. It also can be made roughly quantitative, hut really good analytical A ork requires the presence of an internal gas standard with a convenient mass such as argon (3). The difficulty then is in sample preparation and computation of results, since all significant peak heights must be related to the argon-40 peak. -4method which has been much used in this laboratory is low temperature analytical distillation. It is a time-consuming operation, and, since some of the boron hydrides are very unstable, subject to errors and perhaps impossible in a few cases. Infrared absorption analysis suffers from none of these difficulties. The method is reasonably rapid and the sample is in the spectrophotometer a relatively short time, so that decomposition during analysis is not a serious problem. No standard gas samples need be maintained if the infrared spectrophotometer is kept in good operating condition. The technique identifies the impurities that are present except for those that occur in trace amounts, for which more data would be desirable. EQUIPMENT
A conventional vacuum system consisting of a manifold, liquid nitrogen trap, oil diffusion pump, and mechanical pump was used
