K a 3303.0 Ba 2702.6’ against the concentration of sodium on log-log paper (Figure 1). 4 background correction was made for the sodium line. All experimental samples were arced in triplicate. The reproducibility of the combined chemical and spectrographic analyses was determined by check samples which indicated that the results rrere accurate for sodium to within 20% of the reported value. Figure 1 is a typical plot of the standard curve. For determination of sodium in the barium sulfate, the values obtained from the plot were multiplied by 2 because samples were diluted one half with buffer. Experimental conditions were changed to determine their effect on sodium coprecipitation. ting the ratio of intensities,
PROCEDURE
Effect of Sodium Chloride Concentration. Sodium chloride was added
t o t h e sodium sulfate solution before precipitating the sulfate with barium chloride solution. The increase in sodium coprecipitation with increase in sodium chloride concentration ivas small (Table I). The method of Johnston and coworkers (1. d ) for recovery of sodium sulfate from the barium sulfate precipitate TT as used in one instance. The barium sulfate precipitate originated from the solution containing 100 grams of sodium chloride. The recovered material weighed 13.6 nig. Spcctrographic examination indicated this was principally sodium su1f:tte. Calculated as sodium sulfate this Sives a value of 0.217, sodium, in good agreement with the spectrographic result of 0.24% (Table I). The discrepancy between these results and those of Johnston and ildams (4) remains unexplained.
Table I. Sodium Coprecipitation with Change in Sodium Chloride Concentration
Sodium Chloride Added to Solution, Grams 0
5 10 20 30
50 70 100
Sodium in Barium Sulfate, yo Spectro- Johnston and graphic Adams ( 4 ) 0.09 0.13 0.16 0.14 0.14 0.18
0.21 0.24
0.13 0.28 0.31 0.40 0.51
0.64 0.77
0.74
Effect of Concentration of Barium Chloride Solution. Rate of Addition.
Each solution contained 45 grams of sodium chloride. Changing the barium chloride concentration from 5 t o 10% had no effect on coprecipitation of sodium. Sudden addition of bariuni chloride solution decreases coprecipitation t o half. Effect of Concentration of Sodium Sulfate Solution. T h e sodium sulfate concentration was varied so t h a t barium sulfate precipitates were obtained increasing in rreight by 0.5 gram from 0.5 t o 3 grams. T h e barium chloride concentration was changed for each solution so t h a t 50 ml. would precipit a t e all of the sulfate with a slight excess of barium chloride. Time of addition was 2.5 minutes. KO significant difference in sodium coprecipitation was observed. T i m e of Digestion. Each solution contained 45 grams of sodium chloride. Table I1 compares the results with those of Johnston and Adams (4) , whose solutions contained 5 grams of sodium chloride.
Table II. Change of Sodium Coprecipitation with Time of Digestion
Time of Digestion, Hr. 0 0
0.25 2 3 6
18
20 33
Sodium, % Johnston Spectroand graphic Adams ( 4 ) 0 72 0 50 0 37 0 37
0.16 0 14
o.i5
48
0.24 0 22
0.17
pH of Solution. The p H was varied from 2 t o 6 ; the other experimental conditions were unchanged. S o significant difference in sodium coprecipitation was observed. Fischer and Rhinehammer (5) noted a decrease in n-eight of barium sulfate with decrease in p H and attributed it to coprecipitation of sodium sulfate. However, these results indicate this decrcase is not due to such coprecipitation. LITERATURE CITED
(1) Allen, E. T., Johnston, John, J . Znd. Eng. Chem. 2 , 196 (1910). 121 Crossn-hite, H. AI., Spectrochim. Acta ’ 4, 122 (1951). (3) Fischer, R. B., Rhinehammer, T. B., .&XAL. CHEM.2 5 , 1544 (1953). (4) Johnston, John, .&dams, I. H., J . Ana. Chem. SOC.33, 829 (1911). (5) Riemann, W. H., Hagen, George, IS-D.ENG.CHEN.. ANAL. ED. 14. 150 (1942). (6) Wddbauer, Louis, Gantq E. S., Zbid., 5 , 311 (1933). RECEIVED for review April 4, 1958. Accepted January 12, 1959.
Controlled-Potential Coulometric Determination of Europium W. D. SHULTS Anolytical Chemistry Division, Oak Ridge National laboratory, Oak Ridge, Tenn.
b The europium content of europium oxide can b e precisely determined by controlled-potential coulometry. The electrolytic reduction of europic ions was not satisfactory for direct quantitative titration but was for the preparation of europous ions. Immediate coulometric reoxidation of the europous ions thus produced is the means of quantitative estimation. The method is reasonably rapid and is free from
interference from the usual contaminants.
T
thermal neutron absorption cross section of europium is very large. This, plus the fact that the products of neutron capture by europium possess even larger cross sections, makes it particularly useful as a reactor control rod material ( 2 ) . A precise method HE
for its determination is needed to estimate the exact europium content of commercially available or purified europium oxide. The estimation of europium has not received much attention. Redox titration (IO),in which europic ions are reduced to europous by passage through a Jones reductor and then titrated with standard oxidant, is satisfactory for gram quantities of europium, but is VOL. 31, NO. 6, JUNE 1959
1095
2.0
v
---_
'
1.6-
1.4t
\
1.21 a 1 REDUCTION :
1 i
i b i OXlOATlON :
-801
I
0
I I - 0 . 2 -0.4
I
-0.6
I
-0.8
I -1.0
CONTROL POTENTIAL, volt vs.Ag, AgCl
Figure 1 . Relationship between initial current and electrode control potential for electrolytic reduction of europium(lll) and oxidation of europium(l1) -2.0
not nearly so precise when scaled down for 50- to 100-mg. amounts. For example, duplicate analyses usually agreed only to 2 to 3% when 50 mg. of europium oxide, Eu203, were dissolved in hydrochloric acid and analyzed in this manner in a scaled-down apparatus. Europium can also be determined polarographically (3, 4,8) by reduction a t a dropping mercury electrode, but the precision of this method also is to only 2 to 3%. The well defined and specific waves obtained when europic ions are reduced polarographically a t a dropping mercury cathode and the fact that dipositive europium is easily prepared by electrolytic reduction (6, 11, 14) suggested that controlled-potential coulometric titrimetry might be satisfactory for europium analysis. Europic ions were reduced to europous ions, but with less than lOO~', current efficiency, a t a mercury cathode whose potential was controlled a t -0.8 volt (silver-silver chloride). Various control potentials and supporting electrolytes were studied to obtain adequate current efficiencies, but entirely satisfactory conditions for the direct reduction titration of europic ions could not be established. The oxidation of europous to europic ions at controlled potential did proceed with lOOyocurrent efficiency and is the basis for the method finally evolved. Two electrolyses are required for each aliquot of sample: a reduction of europium(II1) to europium(I1) a t a mercury cathode whose potential is controlled a t -0.8 volt (silver-silver chloride) and an oxidation of europium(I1) to europium(II1) with the electrode potential controlled a t -0.1
1096
ANALYTICAL CHEMISTRY
LI I I 0
1
2
3
4
I \ I 5 6
I 7
I 8
9
I
I
1014
!
12
T I M E , minutes
Figure 2. Relationship between log i for electrolytic reduction of europium(1ll) and oxidation of europium(11) at controlled potential
volt (silver-silver chloride). The quantity of current consumed during oxidation is a measure of europium titrated. This method is rapid and precise, insensitive to acid concentration over a reasonable range, and free from interference from the rare earths and other contaminants found in commercial europium oxide. The procedure as given is arranged so that 5 to 10 nig. of europium are titrated, but smaller or larger quantities were determined successfully. During this investigation, it u-as learned that the application of controlled-potential coulometry to the determination of the rare earth elements is also being studied by Wise and that his results on the determination of europium have been reported (13). APPARATUS AND REAGENTS
The electrical and mechanical apparatus used throughout this study have been described (12). The ORNL electronic controlled-potential coulometric titrator is a simplification of the one described by Booman (1) and its performance and circuitry have been described ( 5 ) . The saucer-shaped glass stirrer used with the titration vessel is positioned a t the mercury-electrolyte interface, its shaft protruding through the cap so that it can be driven by a 1000-p.p.m. Bodine motor (Bodine Electric Co.). The nitrogen gas is thoroughly deoxygenated by passage
through tlvo chromous sulfate scrubbing towers prior to its entry into the cell. Reagents. Euz03 was obtained from Research Chemicals, Inc., and its purity (99.9%) was ascertained by spectrographic analysis. After ignition to constant weight a t 900" C., a weighed portion of this oxide was dissolved by warming in 3 to 4 ml. of 1 t o 10 hydrochloric acid and then diluted t o volume. Appropriate dilutions were made to obtain the desired concentrations. All other reagents were C.P. grade. PROCEDURE
Sample Preparation. Accurately weigh 50 to 100 mg. of Eu2O3powder into a tared 15-ml. platinum crucible. Add 2 to 3 ml. of water and 5 drops of concentrated hydrochloric acid and warm until dissolved. Transfer the solution and rinsings quantitatively to a 10-ml. volumetric flask and dilute to volume. Analysis of Sample. Place approximately 6 ml. of clean mercury, 5 ml. of 0.1M hydrochloric acid, and a 1-ml. aliquot of the sample solution prepared as described in the cell. Make adjustments so t h a t there \vi11 be a reduction a t -0.8 volt (silversilver chloride) and continue the electrolysis until the current reaches and maintains a constant value. (With the present apparatus, this required 7 to 10 minutes.) During this period, deaerate the solution thoroughly with a
stream of pure nitrogen gas. Then stop the titration, short (zero) the integrating capacitor, make adjustments so that a n oxidation at -0.1 volt (silver-silver chloride) will be made, and alloiv the second electrolysis to proceed to completion-until the current falls to 50 pa. Stop this titration and measure the readout voltage. The second titration requires approximately 5 minutes with this apparatus. The weight of europium titrated is calculated by multiplying the readout voltage by the europium coulometric factor. DISCUSSION AND RESULTS
The controlled-potential coulometric reduction of europium(II1) t o europiuni(I1) is complicated because europium(II), a rather powerful reducing agmt, reacts slon-ly with water, resulting in small but intolerable losses in current efficiency and higher than usual background currents. I n attempting to raise the current efficiency to loo%, the use of various acids (hydrochloric, perchloric, and sulfuric) and acid concentrations (0.05 to 2.5M), several salt solutions, and also various buffering systems (formate, acetate, citrate, bisulfate, and barbiturate) as supporting electrolytes for the reduction n-ere investigated. I n no case n-ere the results completely satisfactory but the most promising were obtained with a neutral 2M' magnesium sulfate solution as supporting electrolyte. I n this medium, europous ions are precipitated as insolJble europium sulfate as they are formed. V h e n 10 mg. or more of europium was titrated in this electrolyte, current efficiencies and background currents were satisfactory. but with less tham 10 nig.; current efficiencies fell below 100% because of the increased solubility of europium sulfate. Titrations in magnesium sulfate n-ere also adversely affected by the presence of even small quantities of acid in the sample aliquots and consequently, attention 'rvas directed to the reduction and reoxidation of europium in acid solution. Figure 1 slions the relationship between initial current and electrode control potential when 5 mg. of europium(111) are reduced and 5 mg. of europiuni(11) are oxidized a t a mercury electrode. Potentials more negative than -0.75 volt (silver-silver chloride) can be used for the reduction, and potentials less negative than -0.25 volt (silver-silver chloride) can be used for the oxidation. The values of -0.8 and -0.1 volt hare becm satisfactory in practice. The electrolytic reduction of europiuni(II1) does not proceed with 100% current efficiency nor does the current fall to the usual background value of about 50 pa. n'hen the electrode po-
tential is controlled a t -0.8 volt, the current, depending upon the quantity of europium present, falls to some value near 200 pa. and becomes constant. Europium (111) can be completely reduced to europium(I1) by allowing the electrolysis to proceed until this constant current is reached but the quantity of current consumed at that time is not a true measure of the amount of europium titrated. If the europous ions thus produced are immediately reoxidized, however, n-ith the mercury electrode potential controlled a t -0.1 volt, the quantity of current consumed does become a measure of the amount of europium titrated. This oxidation electrolysis proceeds rapidly to the attainment of a normal background current. A plot of log i 2's. time (Figure 2) graphically s h o w the different characteristics of these tivo titrations because a linear relationship is expected when a single electrode reaction occurs n-ith 100% current efficiency (9). .4liquots of standard europic chloride solutions were analyzed. Two electrolyses were required for each aliquot: a reduction with the mercury (cathode) potential controlled at -0.8 rolt and a coiilometric oxidation with the mercury (anode) potential controlled a t -0.1 volt. The amount of europium titrated was calculated from the current consumed during the oxidation electrolysis. The results of these titrations, listed in Table I, indicate that good precision and accuracy are obtained when 1 to 10 mg. of europium(111) are titrated. Greater quantities than 10 mg. of europium can be titrated without difficulty, but a loss in precision is expected if less than 1 mg. is titrated, because of the small currents associated with such quantities. The procedure given belorv for analysis of Eu208 \?-as designed so that 5 to 10 mg. of europium are titrated. A spectrographic analysis of commercial europium oxide indicated that very small quantities of gadolinium. yttrium, ytterbium. lanthanum, cerium, calcium, aluminum, silicon, and iron may be present. The rare earths reduce only a t potentials much more negative than -0.8 volt ( 7 ) . Calcium, aluminum, and silicon do not interfere. Cerium(1Y) and iron(II1) are reduced to lower valence states during the reduction but do not interfere because neither cerium(II1) nor iron(IT) is oxidized a t -0.1 volt nor do they reduce europium(II1). The influence of the presence of various acids was also checked. Sulfuric acid (and sulfate salts) results in the precipitation of europous sulfate during the reduction step and it is not redissolved upon oxidation. Nitric acid, if present in greater concentration than 0.lM in a I-ml. sample aliquot, interferes by preventing complete reduction
Table 1. Results of Europium Determination b y Reduction and Coulometric Reoxidation
(In each case, a 1-ml. aliquot of standard europic chloride solution was titrated in 5 ml. of 0.1M HC1 by reduction at -0.8 volt and reoxidation a t -0.1 volt. Reference saturated AgC1. Coulometric factor: 62.59 mg. Eu/volt) Relative StandReadout ard Eu Voltage E U (OJida- Found, DeviaPresent, tion, yo Mg. hig. tion) 0.92 0 0150 0.939 0.945 0 0151 0.92 0.27 0.939 0.92 0 0150 0.939 0.92 0 0150 0.945 0.92 0 0151 1.828 1.829 0 0292 1.828 0 0292 1.829 1.821 0.30 0 0291 1.829 1.834 0 0293 1.829 1.828 1.829 0 0292 0 072i 4.550 4.573 4.538 0 0725 4.573 0.17 4.556 0 0728 4.573 4.538 0 0725 4.573 4.556 0 0728 4.573 0 1461 9.144 9.146 9.151 0 1462 9.146 9,157 0.06 0 1463 9.146 9,151 0 1462 9.146 9.144 0 1461 9.146 Table II. Effect of Acid Concentration upon Titration Results
(In each case, a 1-ml, aliquot of europic chloride solution was titrated in 5 ml. of acid of indicated concentration. Coulometric factor: 62.59 mg. Eu/volt) Acidhiolar- Readout Eu, Lfg. Acid ity Voltage Preeent Found HCl 0.01 0.1155 7.617 7.229 0.05 0.1217 7.617 7.617 0.1 0.1217 7.617 7,617 0.5 0.1218 7.617 7 ,623 1.0 0.1222 7.617 7.648 2.0 Kot completed" HC104 0.05 0.1212 7.617 7.586 0.1 0.1219 7.617 7 ,630 0.5 0.1219 7.617 7,630 1.0 0.1219 7.617 7.630 2.5 0.1206a 7.617 7.548 a Reduction electrolysis sluggish
of europiuni(II1) to europium(I1). Hydrochloric and perchloric acids n-ere suitable supporting electrolytes. Aliquots of a europic chloride solution were titrated in hydrochloric and in perchloric acid solutions of various concentrations to determine the influence of acid strength (Table 11). The hj.drochloric acid concentration may vary between 0.05 and 1.OM and the coiicentration of perchloric acid betnwn 0.1 and 1 . O M without deleterious effect upon the titration. ACKNOWLEDGMENT
The author thanks John hfanning and VOL. 31, NO. 6, JUNE 1959
1097
Dave LaValle mho furnished the pure Eu2O3and John Norris who performed the spectrographic analyses. LITERATURE CITED
(1) Booman, G. L., ANAL. CHEM. 29, 213 (1957). (2) Dunning, D. N., Ray, W. E., Nucleonics 10, 88 (1958). (3) Holleck, L., 2. anal. Chem. 116, 161 11939’1. (4f Holieck, L., 2. Elektrochem. 46, 69 (1940).
(5) Kelle
M. T., Jones, H. C., Fisher, D. J., NAL. CHEM. 31, 488 (1959). (6) . _Kleinbera, J., “Unfamiliar Oxidation States ana’ Their Stabilization,” pp. 106-8, University of Kansas Press, Lawrence, Kan., 1950. (7) Kolthoff, I. M., Lingane, J. J., “Polarography ” 2nd ed., pp. 435-41, Interscience, dew York, 1952. (8) Laitinen, H. A., Taebel, W. A., IND.ENG. CHEM., ANAL.ED. 13, 825
(1941). (9) Lingane, J. J., “Electroanalytical Chemistr ” pp. 191-2, Interscience, New Yor%: 1953. (10) McCoy, H. N., J . Am. Chem. SOC. 58, 1577 (1936).
(11) Marsh, J., J . Chem. SOC.1934, 1972. (12) Shults, W. D., Thomason, P. F., ANAL.CHEM.31,492 (1959). (13) Wise, E . N., “Coulometric Analysis of Europium at Controlled Potential,” Paper 21, Southwestern and Rocky
Mountain Division of American Association for the Advancement of Science and the New Mexico Academy of Science, April 30, 1958. (14) Yntema, L. F., J . Am. Chem. SOC. 52, 2782 (1930).
RECEIVEDfor review October 6, 1959. Accepted February 9, 1959.
Determination of Alkanethiols in Hydrocarbons with Silver Ion and Dithizone R. K. KUNKEL, JAMES E. BUCKLEY, and GEORGE GORIN Deparfmenf of Chemistry, Oklahoma Sfate University, Stillwater, Okla.
,Simple and precise methods are described for the titrimetric and the colorimetric determination of alkanethiols in hydrocarbons. Samples (0.01 to 1 mmole) are titrated with silver ion in an ammoniacal alcoholic medium, using ammonium dithizonate as indicator. A sharp color change from orange to red i s observed at the end point. The precision i s to about 0.2%. Smaller samples (0.002 to 0.08 mmole) are analyzed b y adding them to a carbon tetrachloride solution of silver dithizonate in a spectrophotometer cell. After 5 minutes, the absorption due to liberated dithizone i s read a t 61 5 mp, A correction is made for the dilution caused by addition of the sample, and the result i s compared to a standard curve. The precision is to about 2%.
A
LKANETHIOLS occur
to some extent in petroleum, and their removal is an important part of petroleum refining. Because of their objectionable properties, the analytical determination of alkanethiols is of considerable importance. A review of the subject has appeared fairly recently ( 5 ) . Of the reagents employed, silver nitrate has enjoyed the greatest popularity, largely because of its easy availability in satisfactorily pure form. Other reagents which have been proposed must be specially prepared (2, 9 ) . Silver ions react with mercaptans quantitatively t o give silver mercaptides, which usually precipitate. Borgstrom and Reid (3) have utilized the reagent in a titration of the Volhard type, which, however, suffers from various technical difficulties;
1098
ANALYTICAL CHEMISTRY
some have been avoided in a modification proposed by Malisoff and Anding (7). Silver ion has also been used as reagent in electrometric titrations, both potentiometric (8) and amperometric (6). Recently, a coulometric method has been developed (4), which appears to be useful for the routine determination of trace amounts of mercapto compounds in gasoline. While methods that use a visual indicator cannot compete with a coulometric method in routine application, visual indicator methods hold an equally wide margin of advantage when occasional determinations are desired, so that the purchaseandmaintenance of the suitable apparatus is not justified. I n the field of manual operation, a good visual indicator always finds useful application; for the determination of mercaptans with silver ions, the ammonium salt of dithizone is such an indicator. Ammonium dithizonate is pale orange, and it reacts with silver ion in ammoniacal solutions to give a complex, colored deep red. Both forms of the indicator are soluble in isopropyl alcohol, which can also dissolve silver nitrate and is miscible with hgdrocarbons, such as iso-octane. The titration can therefore be carried out in a homogeneous medium. The end point is denoted by a color change from orange to deep red, which is striking and very easily discerned, much more so than the end point in the Volhard-type titrations. Archer (1) has used sodium dithizonate as an indicator for the titration of sulfide with heavy metal ions in aqueous medium. For smaller quantities of alkanethiols,
the determination can be carried out colorimetrically. The thiols are capable of decomposing silver dithizonate, thus liberating free dithizone, which is green and has an absorption maximum a t 615 mp. Therefore, the sample of mercaptan can be added to an excess of silver dithizonate in carbon tetrachloride solution, and the absorption which develops a t 615 mp is measured. Silver dithizonate solution is used as a blank in the determination. Iso-octane was chosen as solvent, because it is typical of the hydrocarbons found in petroleum and can be obtained pure. MATERIALS
Isopropyl alcohol was purified and deaerated by allowing it to react with silver nitrate (5 grams per liter) in direct sunlight for 2 hours, distilling it, and refluxing it a half hour with nitrogen bubbling through. The solvent was stored in a siphon flask and dispensed from it with the aid of compressed nitrogen. Silver nitrate solutions 15-ere prepared by dissolving accurately weighed samples of the reagent in isopropyl alcohol. Most of the data reported lvere obtained with 0.01M solutions, but 0.005M to 0.05144 solutions were employed, a little water being used to aid in the preparation of the more concentrated solutions. .4mmoniacal alcohol solution !?-as prepared by diluting concentrated ammonium hydroxide (30% ammonia) with isopropyl alcohol. Iso-octane, Spectro grade (Brothers Chemical Co., Orange, R’. J.), was treated with, and stored under, nitrogen, as described for isopropyl alcohol. Dithizone indicator was prepared by
