~~
PROCEDURE
Place a 1-gram sample of alloy in a large platinum dish. Add 10 ml. of water and 2 ml. of concentrated sulfuric acid. Slowly add 5 ml. of hydrofluoric acid. After the sample has dissolved, dilute it to about 60 ml., and add 2 nil. uf 30% hydrogen peroxide. The titanium-peroxide complex should impart a golden-yellow color to the solution. Heat slowly on a hot plate until the solution becomes colorless. Maintain the volume of solution a t approximately 60 nil. at all times during the heating period by the addition of mater. Cool, dilute to about 120 ml., and transfer to a flask. If tungsten is present, add 30 ml. of a solution containing 5 grams of tartaric acid. Dilute to a final volume of about 155 nil., and add 1 gram of thioacetamide. After the thioacetamide is dissolved. heat the flask to the temperature of boiling water and maintain that temperature for 60 minutes. Cool to room temperature. at which time the precipitate should settle. The supernatant solution should be clear. Filter through a porous porcelain filtering crucible (Coors KO. 679-3) which previously has been brought to constant tveight at 650" C. K a s h the precipitate with a solution containing 2 nil. of sulfuric acid and 1 gram of tartaric acid diqsolred in 100 ml. of solution. The
Table I.
Allol
A B
c
J>
Analysis
of Titanium Alloys
Spectrographic Analyses, yo ?\Io W A1 Ce Sd 3 11 0 62 0 83 0 00 0 00 2 88 0 88 6 06 0 00 0 00 0 00 1 62 0 00 1 00 1 90 1 98 2 10 0 00 0 00 0 61
tartaric acid may be omitted if tungsten is absent. Ignite the crucible and precipitate a t 650" C. for 60 to 90 minutes. Cool and weigh as molybdenum trioxide. being careful not to expose the crucible to air currents that might displace the powdery precipitate.
Thioacetamide Procedure ?\Io, No. of Std av. cc detns. dev., 5 3 11 6 0 02 2 88 8 0 02 1 00 5 0 05 1 08 6 0 04
partment of Mining and Metallurgy for the titanium alloys used in this research and 0. F. Edwards for the photomicrographs. LITERATURE CITED
(1) Flaschka, H., Jacobljevich. H d n a l . Chim. Acta 4, 482 (1950). (2) JIcSerney, W. N., "Study of the Quantitative Precipitation of AIolybdenum by Thioacetamide," h1.S thesis, University of Kentucky. 1955. (3) Xorwitz, G., Codell, hI ASAL. CHEII 2 5 , 1438 (1953). ( 4 ) Sxift, E. H., Butler, E. A , Zhzrl . 28, 146 (1956). ( 5 ) Yagoda, H., Fales, H. A , J . . i t t i . Chem. Soc. 58, 1494 (1936). ~
RESULTS
Results obtained for the determinntion of molybdenum in four experimental titanium alloys supplied by the Department of Mining and Metallurgy are shown in Table I. The spectrographic analyses for coniparison were performed by a conimercial laboratory. ACKNOWLEDGMENT
The authors nisli to thank J. P. Hanimond and R. E. Swift a t the De-
~
RECEIVEDfor review December 1956. Accepted March 15, 1057.
13,
Constant Cur rent Potentiometric Determination of Manganese CALVIN
0.HUBER
and IRVING SHAlN
Department of Chemistry, University
of Wisconsin, Madison, Wis.
,The technique of constant current potentiometry was applied to two commonly used methods for the determination of manganese. In the Volhard method a significant improvement in accuracy of end point detection is achieved. In the LinganeKarplus method, the titration was as precise and accurate as with the classical potentiometric end point.
A
THE more frequently encountered volumetric methods for the determination of manganese are the Volhard determination ( I O ) , and the Lingane-Karplus method ( 7 ) . The Volhard method involves the oxidation of manganous ion to manganese dioxide with standard permanganate. Although it is possible to obtain quantitative results by using the Fischer (4) modification of the Volhard method, visual end point detection is difficult owing to the JIOXG
1 178
ANALYTICAL CHEMISlRY
presence of brown manganese dioxide formed during the titration. No potentiometric end point method has been reported, although Brann and Clapp (1) investigated the use of a n aniperometric method with one polarized electrode us. a silver-silver chloride electrode. Talsky (9) used a portable spectroscope to detect the first excess permanganate. The Lingane-Karplus method involves the oxidation of manganous ion to manganese(II1) b y permanganate in the presence of a complexing agent, sodium pyrophosphate. Although the end point cannot be detected visually owing to the pink color of the manganese(II1) pyrophosphate complex, excellent results are obtained by following the titration potentiometrically. Both the end point potential and the size of the break at the end point are dependent on pH. Under the best conditions (pH = 6), the potential break is under 300 mv. Thus the titration curve should be plot-
ted for best results. If the titration conditions are reproduced carefully, subsrquent determinations may be made merely by titrating to the predetermined end point potential. Goffart, Nichel, and Pitance ( 6 ) applied a n amperometric method to the detection of this end point, and DUYckaerts ( 3 ) used potentiometry with one polarized electrode. The constant current potentiometric method involves the measurement of the difference in potential between two platinum wire electrodes which are polarized with a constant current. This technique n-as introduced by Dutoit ( 2 ) and was interpreted in terms of voltammetry by Reilley, Cooke, and Furman (8). The mpthod by which the end point is detected can be explained by reference to Figure 1. These experimental current-voltage curves were obtained in solutions containing the same electrolytes a t the same p H as would be present
/I
1
I
I
I
I
+I 0
08
I
I
I
I
I
I
I\
02
06 0 4 VOLTS VS S C E
Figure 1. Experimental current-voltage curves corresponding to solutions present before and after end point in Volhard manganese titration 1. 2. 3. 4
Anodic residual current Oxidation of 10- M manganous ion Reduction of 1 0 - 3 M permanganate ion Cathodic residual current 1
1
t
I
0.8
before and after the end point of a t>yical Yolhard titration. T h e n the t\ro electrodes are polarized Iiy passing a constant currciit through them, the 110tential across the electrodes is indicatecl by the points a t which the currentvoltage curves crosi thc lines rcprescnting the constant current. Rcforc the end point. when only manganous ion< are present, tlie potential differcncc is that betneen points B and D. At thc end point, vhen neither manganous nor permanganate ions are present, tlie pitential difference is indicatcd by point. A and D . Beyond the end point. 11-hen only permanganntc ions arc prestbnt, tlic potential diffrreiice is that b h w n points A and C. An e.rperinicntn1 titration c u r w of voltage against volunir. of standard permaiigaiiatc thewfore ha? the form shown in Figure 2 . Similar current-voltage curl-rs n obtained using tlic soliitioni prcsent hfore and after the mid point in t h r h i gane-Karplus titration. d typical coilstant current potentiometric titration curve for this mrthod is shonn in Fipure 3. The uiqmiiieti ical and rounilrtl shape of tlie titration curves is enu~ctl priiiiarily by the low rates of the c~lcctrode reactions. Surface rractionq on the platinum elcctrodes (6) also contribute to the deviations from throrc.tieally shaped titration curves. Hon CI ('r, these factors do not intcrfere nith the accurate dt~tcctionof the end point. ~ ' J C A
EXPERIMENTAL 0 46 1
27.00
26.00 ML. . 0 2 M KMnO,
Figure 2. Titration curve for constant current potentiometric modification of Volhard-manganese determination
3 3 00
34 0 0
35 0 0 ML
36 00
37 00
O M KMnO.
Figure 3. Titration curve for constant current potentiometric modification of Lingane-Karplus-manganese determination
Apparatus. T h e titrations \ Y ( W carried out in a 300-ml. tall-form beaker. T h e elwtrodes v e r e madc by scaling platinum wire (0.81-nim. tliameter) into soft glass tubing so t h a t a l-cm. l m g t h was exposctl. T h e constant current source K R S a 45-volt battery and 5-, 30-, or 50mrgohni rrsistoi,s connected in srliies \vit h the t IT o platinum ele e t rod cs . This provided 9-, 1.5, or 0.9-pa. polarizing c u r r w t . The potential hetiwen tho c,lrctrodes was measured with a Recknian Model G pH meter. The current-voltage curves used t'o interpret thmr titrat,ions were obtained with a Sargrnt XXI polaroprapli using a c,onwntional rotating platinum cllectrode aswinbly. Extraction. T h r extraction of pyrolusite ores can 1x3 done with hydrochloric acid or with dilute sulfuric acid solutions of hydrogen peroxide or oxalic arid. Interfering exress reduring agents must be removed by boiling to sulfuric acid fumes. Alloys were extracted with 1 t o 1 nitric acid and boiled several minutes with 1 gram of urea t o remove nitrogen oxides. Procedure. VOLHARDMETHOD. Sample sizes were chosen so t h a t 150 t'o 300 mg. of manganese w r c inVOL. 29, NO. 8, AUGUST 1957
* 1179
Table I.
Determination of Manganese in National Bureau of Standards Samples
(Using constant current potentiometric modifications) NBS Sample S o . 66A4 S B S Sample No. 68B (Spiegeleisen)" (Ferromanganese)b Volhard Lingane Volhard Lingane
a
b c
NBS Sample KO.164 (Mn-41 Bronze)" Volhard Lingane
Av. 19.76 19.75 79.99 i9.97 4 678 10.03 St. dev. 1 0 . 0 6 1 0 03 1 0 01 f 0 012 )In 19.7770, C 4.39C/b, P 0.0495, S 0.0217G,and Si 2.26%. \In 79.97%, C 6.77%) P 0.293%, S 0.00970, and Si 0.44%. 11114.687,, Cu 63.76%, Zn 21.89%, A1 6.21%, Fe 2.52%, Pb 0.22%, Sn 0.637,, and Xi 0.045%.
Table II. Determination of Manganese in Commercial Manganese Ore
>In, yo _____ Original Volharda Lingane" Linganeb 61.32 61.35 61.36 61.37 61.34 61.34 61.40 61.35 61.41 61 37 61 34 61 36 61 35 61 42 61 41 61 35 61 42 61 45 61 44 Av 61 37 61 38 61 38 St. dev 1 0 04 f O 04 1 0 03 a Constant current potentiometric modification. Original Lingane-Karplus method (classical potentiometry) cludctl. After extraction, the sample was diluted, 5 grams of sodium acet a t e and 10 grams of zinc sulfate were added, and the saniplc solution was neutralized with zinc oxide paste until no more would dissolve. The solution was then diluted t o volume in a 500-ml. volumetric flask and 100-ml. aliquots were taken from tlic clear supernatant after tmheslight excess of zinc oxide and other insoluble mattersuch as, silicon dioxide and ferric oxidehad settled. The titrations \\-ere carried out' as follo!vs: The aliquot was heated to boiling and about 95% of the permanganate solution required to reach the end point was added without stirring. (The end point, within 2 to 3y0, \vas estimated by s l o ~ l yadding permanganate to a previous aliquot of the hot stirred manganous solution.) After 1 to 2 minutes had been allowed for the reaction to be coniplet'ed, permanganate \vas added until a n apparent end point was reached. Then 1 ml. of acetic acid was added and the titration solution was boilrd for 1 minute. The final end point was found to be that drop which produced a decreased potential, stable for a t least 1 minute. Klien iron was present in relatively large amounts as in ferroalloys, t,he addition of acetic acid was unnecessary. The temperature was maintained ahovc 80" C. throughout the titration process. 1 180
0
ANALYTICAL CHEMISTRY
Slightly before the true end point, potentials indicating presence of permanganate appeared, but were not stable. The change in potential, indicating the end point, n a s a t least 250 mv. and was stable for several minutes. LISGAKC-KARPLUS ~ \ I L T H O U . Sample sizes were chosen so that aliquots would contain 120 to 160 mg. of manganese and were titrated \Tit11 standard 0.02M potassium permanganate. I n the case of manganese-aluminum bronze, the aliquots contained 12 t o 16 mg. of manganese and were titrated with standard 0.002M potasium permangana t e. Aliquots of the acidic sample solutions were added to 200 to 300 nil. of saturated sodium pyrophosphate in a 600-ml. beaker. Critical adjustment of the pH was not required. Equally precise and accurate titrations could be performed in the pH range of 5 to 7 . The end point was taken as the peak of the titration curve. K i t h a little experience, this could be distinguished without taking a series of readings. because of a temporarily lower potential which appeared just a t the end point. The potential decrease, immediately after the end point, was stable and about 100 mi-. per 0.10 nil. RESULTS A N D DISCUSSION
T o test the accuracy of the two methods, several standard samples from the National Bureau of Standards were analyzed for manganese (Table I). As a standard sample of a manganese ore was not available. the results for a commercial ore by these two methods mere conipared with the original LinganeKarplus method (Tablr 11). The relative per cent error is 0.1 or less. I n general, the standard deviation was less for the Lingane-Karplus method than for the Volhard method. The precision obtained for the constant current potentiometric Lingane-Karplus method compares well with that obtained using conventional potentiometry. I n the Yolhard method, substances such as cobalt(I1) and chloride, which are soluhle at the p H of the solution and can be o.tidized by permanganate, interfere. Other common interferences are
4 675
=to 004
removed as the hydrated oxide or basic acetate when the solution to be titrated is neutralized with zinc oxide. Among these are: Ti++++
sn++++
T1+-' -4.35 +
cBi'++ u++
Sn++ Al'+iFe'++
U+++' u 6+
Si++
Cr+++
Cobalt(I1) and chloride, in addition to the above substances, do not interfere in the Lingane-Karplus method. The potentials observed in both titration methods are stable. The large potential observed before the end point in the T'olhard method is probably due to small amounts of manganese(I1) being tied up by the manganese dioxide. These methods are convenient, in that the equipment required is available in any analytical laboratory and no reference electrode is required. The titrations require 10 to 15 minutes to perform. The time required for the determination of manganese by these methods is less than that for the bismuthate method because no filtrations are required. LITERATURE CITED
(1) Brann, B. F., Clapp, Ill. H., J . Am. Chem. SOC.51, 39 (1929). (2) Dutoit, P., Weisse, G. V., J . Chem. Phys. 9, 5i8 (1911). (3) Duyckaerts, G., Anal. Chim. Acta 5,
233 (1951).
( 4 ) Fischer, IT-. I f . , 2.anal. Chem. 48,
751 (1909). (5) Goffart, G., Michel, G., Pitance, T., Anal. Chim. Acta 1, 393 (1947). (6) Kolthoff, I. M., Tanaka, S . , ANAL. CHEJI.26, 632 (1954). ( 7 ) Lingane, J. J., Karplus, R., Ibid., 18, 191 (1946). (8) Reilley, C. N., Cooke, IT. D., Furman, S . H.! Ibid., 23, 1223 (1951). (9) Talsky, G., d n g e w . Chem. 68, 182 (1956). (10) Volhard, J., d47271. Chem. 198, 318 (1879). RECEIVED for review January 7, 1957. .4ccepted March 21, 1957. This work was supported in part by the Wisconsin A41umniResearch Foundation.
