I n summary, the criteria used to establish the most accurate metallic iron values in reduced iron ore were material balances near 100% and calculated oxygen content which agreed with the direct determination. Based on this, it has been shown that the bromine-methanol procedure with 15-minute reflux time yields the most accurate results. The mercury(I1) chloride procedure tends t o be biased on the low side a t all levels, while the silver thiocyanate method yields consistently high
results, which worsen at lower metallic iron levels. The bromine methanol method has been in routine use in several laboratories for over three years. During that time many different types of ores in various degrees of reduction have been successfully analyzed. RECEIVED for review August 6, 1970. Accepted October 22, 1970.
Amperometric Titration of Calcium and Magnesium Using the Lead Dioxide Electrode C. 0. Huber, K. Dahnke,’ and F. Hinz’ Department of Chemistry and Laboratory f o r Surface Studies, Uniaersity of Wisconsin-Milwaukee, THEDETERMINATION of calcium and magnesium in the same solution is a n important analytical chemistry problem because of the occurrence of both of these elements in drinking water, body fluids, etc. The concentrations encountered and the accuracy required are such that titration is often the most suitable method for analysis. Attempts have been made in the past to titrate the calcium specifically with EDTA (ethylenediaminetetraacetic acid) while preventing magnesium reaction by precipitating magnesium hydroxide at p H 12. Convenient and specific titration of calcium in the presence of magnesium has been demonstrated by Hildebrand, Schmid, and Reilley ( I , 2) using disodium EGTA (ethylene glycol bis-paminoethyl ether-N,N’-tetraacetic acid). When a suitable end point indicating system is available, the magnesium can subsequently be titrated using EDTA. Monnier and Roueche (3) performed such titrations amperometrically with a stationary mercury electrode at pH 10.5. Ordinary ammonia buffer systems are not permissible in their method because of mercury amino complex formations. Here we report a n amperometric titration method based on the anodic response of the lead dioxide electrode to EDTA and to EGTA. An ordinary ammonia-ammonium buffer system is used. The titration is rapid and convenient. Titrant concentration as low as millimolar can be used. Titration curve shapes are explained with reference to previously reported properties of the lead dioxide-solution interface. EXPERIMENTAL
Apparatus. Except for the working electrode, conventional solid electrode amperometric apparatus was used. The cathode was a saturated calomel reference electrode connected by a salt bridge. The titration solutions were stirred with an electric motor driven magnetic stirrer. The working electrode was PbOe plated onto approximately 12 X 1 mm platinum wire sealed in glass. The plating was done from 0.1M Pb(N03)2 in 0.5M H N 0 3 . A suitable plate was obtained by plating at 6.0 mA for 30 minutes. The plating solution was stirred with a magnetic stirrer. Treatment of the platinum wire with concentrated HC1 previous to Present address, Department of Chemistry, Purdue University, Lafayette, Ind. 47907 (1) G. P. Hildebrand and C. N. Reilley, ANAL. CHEM., 29, 258 (1957). (2) R . W. Schmid and C. N. Reilley, ibid., p 264. (3) D. Monnier and A. Roueche, Helc. Clrim. Acta., 41, 103 (1964). 152
Milwaukee, Wis. 53201
plating the lead dioxide removed old PbO2 and also improved the apparent quality of the P b 0 2plate. A stable and sensitive P b 0 2 layer appeared almost black. Any exposed areas of platinum on the electrode resulted in spurious currents and could not be tolerated. Once prepared, the anode indicator electrode could be used for at least 20 titrations and often remained suitable for many more. When not in use, the electrode was stored in distilled water. Reagents. Standard solutions of 0.01M EDTA and 0.01M C a were prepared using reagent grade disodium EDTA salt and calcium carbonate. In order to remove any EDTA impurity in the reagent grade EGTA, it was recrystallized by acidifying an alkaline solution containing a small amount of magnesium. This was necessary because initial results showed persistently high results for calcium and low results for magnesium such as would be caused by EDTA as an impurity in the EGTA. Magnesium and EGTA solutions were standardized by titration with the standard EDTA and calcium solutions. The buffer solution was prepared to be one-tenth molar in both reagent grade ammonia and ammonium nitrate and then adjusted to p H 11 using sodium hydroxide. Standard serums were obtained commercially (Dade Reagents, Inc., Miami, Fla.) and used as purchased. Solutions were stored in polyethylene bottles. Procedure, A 10.0-ml aliquot of the sample solution is added to 10 ml of 0.1M buffer solution in a 50-ml beaker which serves as the titration cell. After positioning the electrode and salt bridge, the applied voltage is adjusted to +1.2 volts as. SCE. Titrant concentration is either 0.01M or 0.001 M depending on sample concentration. Distinct changes in slope of the current us. volume of titrant plots mark the equivalence points. RESULTS AND DISCUSSION
Titration Curves. A typical titration curve for 0.001M titrant appears in Figure 1. The calcium end point is marked by the first stable drop in the anodic current. The titrant is then changed to EDTA, and the magnesium is titrated to the linearly extrapolated end point as shown. Only a small excess of EGTA should be added after the first end point, less than 0.5 ml of 10-3M EGTA if possible. Excess EGTA does not ruin the stoichiometry of the magnesium titration, but a larger excess tends to “round” the second titration equivalence point region because of conversion of MgEGTA to MgEDTA. Titrations performed at various applied voltages between $0.8 and +1.4 volts as. SCE give the same general shape of the titration curve shown in Figure 1. The choice of 1.2 volts
ANALYTICAL CHEMISTRY, VOL. 43, NO. 1, JANUARY 1971
Table I. Titration Results
Sample 10-zM Ca and Mg 10-3M Ca and Mg
Hard water Std serum
Present 408 ppm Ca
Found 407 ppm Ca
240 ppm Mg 40.2 ppm Ca 24.0 ppm Mg 3 5 . 0 ppm Ca 11.2ppmMg 161 ppm Ca 50 ppm Mg
239 ppm Mg 40.0 ppm Ca 23.7 pprn Mg 34.2 ppm Ca 10.9 ppm Mg 159 pprn Ca 53 ppm Mg
is based o n the observation that changes in current were slightly larger at this potential. Titration Results. Results for titrations of various samples containing calcium and magnesium are shown in Table I. Hard water and blood serum can be titrated without previous separations. The accuracy and precision compare well with other methods at these concentration levels. Relatively large amounts of phosphates and heavy metals would be expected to interfere. Results with the standard serums indicate, however, that for these samples, chloride, iron, and phosphate in concentrations at least equal to those for calcium and magnesium do not interfere. Magnesium solutions titrated with EGTA behaved just as blank solutions. Also, calcium end points occur at the equivalence points even in the presence of magnesium at ten times the calcium concentration. These observations confirm that the calcium equivalence point indication is unaffected by the presence of magnesium in solution. Electrode Response Mechanism. At the lead dioxide electrode EDTA and EGTA yield anodic currents at potentials more positive than about 0.8 V cs. SCE. The mechanism for this anodic current has been discussed previously ( 4 ) . Part of that mechanism involves slow, but continuous reduction of the electrode material to the lead two state by the solvent system. Reoxidation of at least a portion of the divalent lead accounts for the relatively large anodic background current observed before the first equivalence point in Figure 1. First End Point Signal. If titration is continued much beyond the first end point with EGTA, the initial drop in anodic current shown in Figure 1 reaches a minimum and then increasing anodic current is observed. Thus a limited amount of EGTA produces a descreasing current after which the increasing anodic current mechanism referred to earlier applies. In fact, when 0.01M rather than 0.001M EGTA is used, this initial decrease just after the end point is not resolved and the anodic current begins to increase immediately giving the same general appearance for the first as for the second end point. This limited decrease in anodic current at the first end point is attributed to chelation of divalent lead species in the reaction layer of the electrode by the excess chelating agent occurring after the equivalence point has been reached. This chelation prevents reoxidation of the divalent lead. The fact that this decrease is limited to a small amount of free EGTA indicates that only a restricted amount of divalent lead in the reaction layer is involved. The divalent lead would otherwise contribute to the background current as discussed above. One practical result of these observations is that the titrant con(4) D. R. Tallant and C. 0. Huber, J . Electroanal. Chem., 18, 413 (1968).
Est std dev
No. of
titrations 5 5 5 5 5
1 . 7 ppm 1.4 0.1 0.3 0.8 0.3 2.2 2.0
5
4 4
YA
.
~
_L-_
5 M
L
IO
T
I
T
R
b
N
T
Figure 1. Titration curve
centration should be low enough so that the decrease in current is clearly observed or otherwise high enough so that the current clearly increases after the equivalence point. The effect itself should not be dependent o n electrode area, therefore the choice of either 0.001M o r 0.01M titrant used here can be recommended for other laboratories. Apparently at certain titrant concentrations between these values, a somewhat ambiguous first end point would be obtained. Electrode reactions in the vicinity of the end point were further studied using blank titrations. For both 0.001M EDTA and 0.001M EGTA, the sharp drop in current shown for the first equivalence point in Figure 1 is also observed upon the first additions of ligand to a blank solution. Thus, there is no measurable discrepancy between equivalence point and end point. This also shows that when used separately, an initial excess of either ligand first depletes lower oxidation state lead species in the reaction layer. In both cases after about 0.25 ml of 0.001M titrant had been added to the blank solution, the decreases in current were superseded by increases in current as described above. Second End Point Signal. At the second equivalence point, the lower oxidation state lead species have already been depleted, so that only the increased anodic current can occur when the end point has been reached and excess EDTA is present. Electrode Preparation and Storage. Variation of the current density when plating the lead dioxide electrode demonstrated that electrode sensitivity to free titrant increases with decreasing plating current density. Presumably lead dioxide layers deposited with lower current density contain more active sites in the lead dioxide lattice. The plating procedure adopted was based on obtaining adequate sensitivity without inconveniently long plating time. Once prepared, storage of the electrode in distilled water
ANALYTICAL CHEMISTRY, VOL. 43, NO. 1, JANUARY 1971
153
results in a higher sensitivity than when stored in air. Such dependence o n environment is not unexpected for an active electrode surface. The results here suggest that the lead dioxide electrode is potentially applicable to the monitoring and investigation of a wide variety of titration reactions.
RECEIVED for review May 26, 1969. Accepted October 27, 1970. Presented at Division of Analytical Chemistry, 156th Meeting, ACS, Atlantic City, N. J., September 1968. Work supported in part by the National Science Foundation-Undergraduate Research Participation Program (GE-6510, GY3044).
Determination of Nitrogen in Vanadium Nitrides and Uranium Nitride (UN) Dean H. Bollman and D. M. Mortimore Bureau of Mines, United States Department of the Interior, Box 70, Albany, Ore.
FORTHE KJELDAHLdetermination of nitrogen in refractory nitrides, one of the most exasperating problems is the inability in some cases to dissolve appropriate amounts of sample without loss of nitrogen. Silve Kallmann er a!. ( I ) have shown that a hydrofluoric acid-phosphoric acid-potassium dichromate combination will dissolve various refractory metals rather rapidly. The method is also applicable to most carbides and nitrides. However, no mention was made of vanadium nitride, which is particularly troublesome with regard to the problem mentioned above. Other workers (2, 3) have gone to modifications of the Dumas method for the determination of nitrogen in uranium nitride and vanadium nitride. Morgenthaler and Menichelli ( 4 ) used a gasometric determination requiring the use of chlorine. E. T. Turkdogan and E. M. Fenn (5) used the Dumas method for determining the amount of nitrogen in vanadium nitrides when they were investigating the variation of composition with temperature of vanadium nitrides under one atmosphere of nitrogen. At 900 “C, a nitride contained 2 0 . 5 z nitrogen, but at 1500 O C it had decreased to 17.1 nitrogen. They reported that the vanadium nitrides have distinct composition ranges, the beta ranging from VNo,;, to VNo.as,and the gamma from VN0.71to vN1.0~.The fact that the vanadium nitrides are not stoichiometric may be a factor in the nitrogen losses occurring during the regular Kjeldahl determination. Meyer, Parry, and Davis (6) successfully used tube furnace ignition and gas chromatographic determination of the nitrogen for determining nitrogen in various refractory nitrides, including vanadium nitride. (1) S. Kallmann, E. W. Hobart, H. K. Oberthin, and W. C. Brienza, Jr., ANAL.CHEM.,40, 332-335 (1968). (2) Research Group Rep. AERE-R 4644; Scientific Administra-
tion Office, Atomic Energy Research Establishment, Harwell, Didcot, Berkshire, England, August 1964. (3) 0. H. Kriege, AEC Rep, LA-2306, Los Alamos, N. M. Aug.
1959, p 38. (4) L. P. Morgenthaler and R. P. Menichelli, ANAL.CHEM., 37, 570-571 (1965). (5) E. T. Turkdogan and E. M. Fenn, J. Iron Steel Inst., 181, 343344 (1955). (6) R. A. Meyer, E. P. Parry, and J. H. Davis, ANAL.CHEM.,39, 1321-1323 (1967). 154
For uranium, UOn, and uranium nitrides, Joan Lathouse, F. E. Huber, Jr., and D. L. Chase (7) and others (8, 9) have shown that loss of nitrogen can occur in the dissolution process for Kjeldahl determinations. Cupric selenate and hydrofluosilicic acid used in the dissolving process and D e Varda’s alloy used during distillation helped in some cases to prevent losses. However, Sinclair, Davies, and Melhuish (IO) reported results which show that cupric selenate promotes loss of nitrogen when uranium-plutonium dioxides were dissolved in a mixture of phosphoric and sulfuric acids. Turovtseva and Kunin (11) suggest that insoluble residues be fused with potassium sulfate crystals and a mixture of phosphoric and sulfuric acid. Our experience with bisulfate fusion of vanadium nitride has been that temperatures high enough to decompose the nitrides invariably led to nearly complete loss of nitrogen. To obtain accurate and simplified methods for nitrogen in vanadium nitrides and UN, various procedures were attempted. Kjeldahl procedures for vanadium nitrides and uranium nitride are given below. EXPERIMENTAL
Apparatus used were two semimicro Kjeldahl distillation flasks, 270-ml capacity; a n Erlenmeyer flask, 4-liter, thickwalled, with Nichrome wire heating assembly mounted through a rubber stopper as a steam generator; a n acid digestion bomb with Teflon (Du Pont) cup; and other conventional laboratory glassware and equipment. Reagents were sodium hydroxide solution40 NaOH with HsO, cool; sulfuric acid titrating solution-approximately 0.14N; boric acid solution, 1 % with methyl red indicator. Combine 0.090 gram methyl red in 90 ml of ethyl alcohol, with 90 grams boric acid in 9 liters of H20. Make up and store in a 2 1 / 2 gallon solution bottle. The boric acid should be com(7) Joan Lathouse, F. E. Huber Jr., and D. L. Chase, ANAL. CHEM.,31, 1606-1607 (1959). (8) Production Group Rep. 374(S), United Kingdom Atomic
Energy Authority, Risley, Warrington, Lancashire, England, 1962. (9) Harlan J. Anderson and James C. Langford, ANAL.CHEM., 35, 1093-1094 (1963). (10) V. M. Sinclair, W. Davies, and K. R. Melhuish, Tuluntu, 12, 841-845 (1965). (11) 2. M. Turovtseva and L. L. Kunin, “Analysis of Gases in Metals,” Consultants Bureau, New York, 211-212 (1961).
ANALYTICAL CHEMISTRY, VOL. 43, NO. 1, JANUARY 1971
