spectrophotometric method (Table 11). The accuracy of the polarographic method was satisfactory, as is evident from Table 11.
Table II.
Sample
RESULTS
SO.
Samples of thoria, thorium metal, and thorium nitrate were carried through the analytical procedure (Table 11). Some of the samples were also analyzed iiy a spectrophotometric method (11) after zirconium had been removed b y anion exchange from 1231 HC1, and agreement with the polarographic results was excellent. Even Specpure thorium compounds were found to contain as much as 150 p.p.m. of zirconium, which cauwd positive errors in the spectrophotomchc method unless removed.
1
2
3
Determination of Aluminum in Thorium Compounds
A1 Found Polarographically, P.P.M. -I
(1) Burelbach, J. P., March, R. J., U. S. -4tomic Energy Conim. Rept. ANL5240 (1953). ( 2 ) Cooney, B. A., Ph.D. thesis, Duke
5.00
8, 10
a
ACKNOWLEDGMENT
The author is grateful to Yvonne Farrar for help in obtaining some of the experimental results.
Spike Recovered aril Found Spike from SpectrophotoAdded t o Sample metrically Sample (pg. of Al) (pg. of Al) (11), P.P.M. 5.00 5.1 3
-
0
2E -0.1
(r
0.4M
t-
//
II
0 1
0
-0.j
-0.2
-0.3 E vs S.C.E.
- 0.4
Figure 1. Variation of formal redox Dotential with hvdr&hloric acid concentration
calibration of this instrument are available (3). The conventional titration cell used a mechanically stirred mercury cathode and a saturated calomel reference electrode (10). Reagents. A standard solution of Sb(II1) was made b y dissolving (and diluting to volume) a known weight of analytical reagent grade SbC1, in 6 M HCl. This solution was prepared to have a concentration of 10.34 mg. of S b per ml. and was standardized by potentiometric titration with K2Cr20,. A standard solution of Sb(V) was prepared by diluting an aliquot of SbCls (99.8% purity) to volume with 6M HC1. This solution was prepared to contain 9.53 mg. of Sb per ml.; it was standardized iodometrically. A standard solution of Sb was prepared by dissolving a known weight of lump Sb metal (99.9yo purity) in tartaric and nitric acids and diluting to volume. Procedure. I, A . Chemical Reduction of Sb(V). Pipet a sample test portion which contains 2 t o 8 mg. of S b into a 30-ml. beaker and dilute to 5 ml. with 2-11‘ HC1. Add 50 to 100 pl. of 85% N2H4.H20. Warm the solution at incipient boiling 10 to 15 minutes on a hot plate. Cool and transfer into the titration vessel, using as wash 0.5M HCl and 5 ml. of 0.8M H2C4H406. I, B. Titration of Sb(II1). To the supporting electrolyte solution composed of -1iZI HC1 and 0.4M H2C4H4o6 and containing 2 to 8 mg. of Sb(III), add 8 ml. of Hg and deaerate the solution 5 minutes with a brisk stream of He. Prereduce the solution at -0.08 volt us. S.C.E. under a He blanket until the background current decreases to 100 pa, and zero the integrator. Reduce Sb(II1) to Sb(Hg) a t -0.28 volt us. S.C.E. until the current decreases to 100 pa. Using the integrated current consumed during this reduction, calculate the amount of Sb(II1) titrated via Faraday’s law with n = 3. 11. Titration of Sb(V) and Sb(V), Sb(II1) Mixtures. Place 10 ml. of 0.4M H2C4H40a-6MHC1 solution and 8 ml. of H g in the titration vessel. Prereduce the electrolyte at -0.35 volt os. S.C.E. until the current de-
500
ANALYTICAL CHEMISTRY
I
Figure 2. Electrolytic reduction of antimony(V) by two-step process in 6M hydrochloric acid
creases to 100 pa. Pipet into the prereduced solution a sample test portion containing approximately 5 mg. of Sb and deaerate the solution 5 minutes with a brisk stream of He gas. Reduce Sb(V) to Sb(II1) a t -0.21 volt us. S.C.E. until the current decreases to 100 Ma. The integrated current consumed is a measure of Sb(V) content via Faraday’s law with n = 2. Xext, reduce Sb(II1) to Sb(Hg) at -0.35 volt us. S.C.E. until the current decreases to 100 ?a. The integrated current consumed is a measure of total Sb content via Faraday’s law with n = 3. Antimony(II1) content is determined by difference. DISCUSSION A N D RESULTS
Curves are given in Figures 1 and 2 which relate readout voltage, a measure of the quantity of current consumed during electrolysis, with controlled electrode potential when antimony \vas reduced electrolytically in several supporting electrolyte solutions. Data for these curves were obtained by making the reduction in the usual manner except that periodically throughout the electrolysis the potential was adjusted to a value which caused cessation of current flow. The readout voltage (proportional to current consumed or extent of reaction) and electrode potential were noted and the electrolysis was then continued. Curves plotted from a series of such points are especially useful in comparing supporting electrolyte solutions and in establishing optimum electrode potentials because they relate extent of reaction with electrode potential under actual titration conditions. Antimony(II1). Antimony (111) in HC1 medium can be electrolytically reduced a t a mercury cathode b y a reversible 3-electron process b u t as the solid curves in Figure 1 indicate, the formal redox potential of the Sb(II1) e Sb(Hg) couple varies considerably with HCl concentration. Consequently, while Sb(II1) can be titrated in an HCl medium, the concentration of HCl in the supporting elec-
trolyte solution must be knolm and held constant to establish with certainty the proper electrode potential for reduction. Furthermore, reductions in strong HCl medium are characterized by undesirably large background currents and evtended titration times with resulting high results. These difficulties can be eliminated by including tartaric acid in the supporting electrolyte solution. As shorrn by the dashed curves in Figure 1, the formal redox potential of the Sb(II1) e Sb(Hg) couple changes by only 40 mv. or so when HCl concentration in the supporting electrolyte solution is increased from 0.3 to 2.5111, provided tartaric acid is present. Tartaric acid concentration was constant a t 0.431 during these electrolyses but the tartaric acid Concentration can vary from 0.1 to 0.8M with little effect upon the Sb(II1) e Sb(Hg) potential. Thus, the optimum supporting electrolyte solution for coulometric reduction of Sb(II1) to Sb(Hg) is 0.4X H2C4H406 that contains approximately 1M HC1. Titrations can be made a t -0.28 volt us. S.C.E. in this medium when the HCl concentration does not exceed 2.511f, background currents and titration times are reasonably small, and the precision and accuracy are improved. Coulometric oxidation of the Sb(Hg) resulting from reduction can also be made a t -0.08 volt us. S.C.E. with good precision, but the results are generally low by 0.5 to lY0. It is postulated that the formation and loss of a small amount of stibine during reduction accounts for the low oxidation results. This oxidation step is not required in analysis of Sb(II1) samples. The data in Table I were obtained by analysis of standard antimonous chloride solutions by the procedure for coulometric titration of Sb(II1) ; they are indicative of the precision and accuracy that can be obtained by that procedure. Antimony(V). Total antimony, if present as Sb(V), can be determined
by this procedure only if t h e Sb(V) is first reduced t o Sb(II1). A large number of reducing agents were considered for this purpose. Of these, hydrazine hydrate proved t o be most satisfactory because i t accomplishes quantitative reduction of Sb(V) t o Sb(II1) but not to the metal and because any excess reductant does not interfere in subsequent electrolyses. Procedure I, A gives the method used for chemical reduction prior to coulometric titration by Procedure I, B. D a t a given in Table I1 were obtained by analysis of a standard antimonic chloride solution by procedures I, A and B; they indicate that the precision and accuracy of the combined procedure are not adversely affected by the chemical reduction treatment. Table I1 also gives the precision obtained when a KSbFs solution was analyzed by procedures I, A and B. It is not possible to report accuracy because this fluoride solution could not be standardized by any of several other methods (potentiometric and iodometric titration and sulfide precipitation). The prccisiori is acceptable. -4ntimonj (V) can be made to undergo a two-step electrolytic reduction by use of a medium containing strong HC1, as the curves in Figure 2 indicate. Furthermore, by judicious choice of electrode potentials, it is possible to determine Sb(T’) by electrolytic reduction to Sb(II1) and then determine total antimony by the Sb(II1) to Sb(Hg) reduction. Antiniony(II1) content obviously is determinable by difference. As the curves in Figure 2 show, addition of tartaric acid to this medium has little effect upon the rcdov potentials of the Sb(V) $ Sb(II1) and Sb(II1) e Sb(Hg) couples. Inclusion of tartaric acid is desirable, however, because it results in smaller reagent blanks and background currents, and in improved accuracy, particularly when the amount of antimony titrated is small. Background currents in strong HC1, even when tartaric acid is present, tend to be large, e.g., 100 to 200 pa. in 6.11 HCl. Attempts to decrease the magnitude of the background current by substitution of LiCl for part or all of the €IC1 in the supporting electrolyte solution were unsuccessful; both high acidity and high chloride content are necessary for the electrolytic reduction of Sb(V) to Sb(II1) to be satisfactory. Kevertheless, good results can be obtained in the analysis of Sb(V) and of Sb(V), Sb(II1) mixtures if the supporting electrolyte solution alone is pretitrated to background as called for by procedure 11. Table I11 presents the results obtained when standard SbC16 and SbC16, SbC13 mixtures in varying proportions were analyzed by procedure 11. Procedure I1 gives results that are slightly
Table 1.
Results of Analysis of Standard Antimony(ll1) Solutions by Procedure I, 6
n
Taken
12 13
2.59 5.17
Table II.
Sb(III), Mg. Found Red. Oxid. 2.59 5.17
u. c7, I
2.57 5.14
I -
Red.
Oxid.
0.4 0.3
0.4 0.3
Results of Analysis of Antimony(V) b y Procedures I, A and
B
I
Species
n
Taken
SbW KSbFs
15 10
4.76 8
Table 111.
Sb(V), Mg. Found Red. Oxid. 4.76 8.06
u., 41, I -
4.72 8.01
Red.
Oxid.
0.2 0.3
0.2 0.3
Results of Analysis of Antimony(V) and Antimony(V), (111) Mixtures
(Aliquots of SbCls and SbCla standard solutions were analyzed by procedure 11) Sb(V) 6 6 4 4 5 5
2.38 4.76 2.38 4.76 2.38
?Tone
Taken, Mg. Sb(1II) None None 2.07 2.07 5.17 5.17
Total
Sb(V)
2.38 4.76 4.45 6.83 7.55 5.17
2.38 4.75 2.39 4.76 2.39
Found, Mg. Sb(II1) hTone None 2.08 2.08 5.18 5.19
None Precision, %: Sb(V), u = 0 . 4 Sb total, u = 0 . 3 Sb(III), u = 0 . 7
less precise than those obtained by procedure I, B because of the large concentration of HC1 required in the supporting electrolyte and particularly because Sb(II1) results are obtained by difference. A standard solution, prepared from high purity antimony metal by dissolution in tartaric and nitric acids t o contain 27.49 mg. of Sb per ml., was also analyzed by procedure I1 for total antimony and 27.48 mg. of Sb per ml. was found. The oxidation state distribution in this solution was uncertain because of the manner of dissolution, thus the solution was not analyzed for oxidation state distribution. Interferences. T h e effects of several anions on t h e results obtained b y procedure I were studied a n d are presented in Table IV. Reasonable amounts of nitrate, sulfate, and fluoride d o not interfere in t h e titration. However, when nitrate is present, 5 drops of a saturated sulfamic acid solution should be added t o destroy traces of nitrites. Table V tabulates data obtained during a study of the effects of several
Total 2.39 4.76 4.47 6.84 7.57 5.19
Table IV. Effect of Anions on Coulometric Titration of Antimony by Procedure I, B
[iintimony(III) taken in each titration, 5.17 mg.]
ilnion Added,
Sb(II1) Found,
Mg. 44 NO3-a 88 KO3258 NOS416 NOa78 504--2 156 SO;-2 312 10 F20 F40 F60 F80 F-
Mg. 5.17 5.20 5.18 5.22 5.17 5.18 5.18 5.17 5.17 5.15 5.22 5.24
Error,
70
0.0 +0.6 $0.2 +1.0
0.0
$0.2 $0.2 0.0 0.0 -0.4 +l .o $1.4
a Five drops of saturated solution of sulfamic acid added t o all titrations which contained NOa- t o destroy traces of nitrites.
cations on the coulometric titration of antimony(V). The effects of these potential interferences were studied by VOL. 34. NO.
4, APRIL 1962
501
procedure I1 because it appeared to be more widely applicable than procedure I. This investigation indicated that Cu(I1) cannot be tolerated without interference. Copper is reduced with antimony at both -0.21 and -0.35 volt us. S.C.E., and it is also reoxidized with the antimony. The presence of moderate amounts of As(V) and Pb(I1) apparently do not affect this titration to any large extent although there appears to be a small negative error in the reduction of Sb(V) to Sb(II1) a t -0.21 volt us. S.C.E. when Pb(I1) is present. Tin interferes in the reduction steps when present either in the stannic or the stannous state, but total antimony can be determined by reoxidation with very good precision. The interferences of Fe(III), Ni(II), and U(V1) are eliminated by reduction a t -0.21 volt us. S.C.E.; total antimony is then determined by reduction a t -0.35 volt us. S.C.E.Bismuth(V) does not interfere in the reduction of Sb(V) to Sb(II1) at -0.21 volt us. S.C.E.However, it is reduced with Sb(II1) at -0.35 volt us. S.C.E., so that only Sb(V) can be determined in the presence of Bi without interference.
LITERATURE CITED
Table V. Effect of Cations on Coulometric Titration of Antimony by Procedure II
[Each value is average of a t least 3 trials. Antimony(V) taken in each titration, 4.76 mg] Cation Sb Found, Error, 071 Added, Mg. blg. /O 2 . 5 $effi 4.730 -0.6 4.77b $0.2 3 Bif6 4.744 -0.4 $10.8 2 . 5 cu+2 5.3Zb 4,776 +0.2 2.5 Fe+3 +0.4 4.78b 5 Fe+3 5 Ni+2 +0.6 4.79b -1.3 4.70a 2 . 5 Pbf2 +0.4 4.7@ 5 Pb+2 -0.8 4.773 4.79b +0.6 4.7% +0.4 2.5 Sn+2 +0.8 4 . 8OC 5 Sn+2 +0.4 4.7P 2.5 Sn+4 +1.1 4.8lC 5 Snt4 $0.2 4.776 3 u+o $0.4 4. 78b 6 U+6 a Value obtained by reduction a t -0.21 volt us. S.C.E. *Value obtained by reduction a t -0.35 volt us. S.C.E. c Value obtained by oxidation at -0.21 volt us. S.C.E.
(1) Diehl, H., “Electrochemical Analysis
with Graded Cathode Potential Control,” p. 45, G. Frederick Smith Chemical Co., Columbus, Ohio, 1948. (2) Hayakawa, H., Bunseki Kagaku 7, 360 (1958). (3) Jones, H. C., “Automatic Coulometric Titrator, ORNL Model Q-2005, Electronic, Controlled-Potential,” Methods 1 003029 and 9 003029 TID-7015,See. 1, -4ugust 17, 1959. ( 4 ) Kelley, M. T., Jones, H. C., Fisher, D. J., ANAL. CHEY.31, 488, 956 (1959). (5) Kolthoff, I. M., Lingane, J. J., “Polarography,” 2nd ed., pp. 545-50, Vol. 11, Interscience, New York, 1952. (6) Lingane, J. J., Nishida, Fumio, J . Am. Chem. SOC.69, 530 (1947). 17) Norwitz,, G.,, -4XAL. CHEX 23, 386 (1951). (8) Reynolds, S. A., U.S.Atomic Energy Comm. Rept. ORNL-1557(1957). (9) Schleicher, A., Toussaint, L., Chem. Ztg. 49,645 (1925). (10) Shults, M7. D., “Uranium, Automatic Controlled-Potential Coulometric Titration Method,” Methods 1 219225 and 9 00719225; TID-7015,Sec. 1, January 29, 1960. (11) Tanaka, M., Bunseki Kagaku 7, 296, 631 (1958). RECEIVED for review November 17, 1961. Accepted January 17, 1962. Fifth Conference on Analytical Chemistry in Nuclear Reactor Technology, Gatlinburg, Tenn., October 12, 1961.
Spectrofluorometer Calibration in the UIt raviolet Region C. A. PARKER Admiralty Materials Laboratory, Holfon Heafh, Poole, Dorsef, England
b A simple method is described for calibrating a fluorescence spectrometer so that corrected fluorescence emission spectra in the ultraviolet region can be determined. The method makes use of a fluorescent screen monitor which was originally designed to allow direct recording of corrected fluorescence excitation spectra. Corrected fluorescence emission spectra and relative fluorescence efficiencies of anthracene, naphthalene, phenol, and benzene are presented.
M
SPECTROFLUOROMETERS record “apparent” excitation and emission spectra which are a function of the particular instrument used. Such apparent spectra are often grossly distorted versions of the true spectra, and before they can be compared with results obtained with other instruments, they must be corrected to give the true spectra. Methods for determining true fluorescence excitation spectra throughout the visible and quartz
502
OST
0
ANALYTICAL CHEMISTRY
ultraviolet region have been described (5, ’)I but a convenient method for correcting fluorescence emission spectra in the ultraviolet region has not previously been reported. The true fluorescence emission spectrum of a solution is a plot of fluorescence intensity, measured in relative quanta per unit frequency interval, against frequency ( 7 ) . When the fluorescence monochromator is scanned a t constant slit width and constant detector sensitivity, the curve obtained is the apparent eniission spectrum. T o determine the true spectrum, the apparent curve has to be corrected for changes of the sensitivity of the multiplier phototube, the band width of the monochromator, and the transmission of the monochromator with frequency. Thus, if dQ/dv represents the fluorescence intensity a t any frequency v , the observed multiplier phototube output, A , , which corresponds to the apparent emission spectrum, is given by: AD =
yz)
PYBYLY =
rz)
SU
where
P,
output per quantum of the multiplier phototube at frequency Y B,, = band rvidth in freauencv ~. . ” units at frequency v L , = fraction of light transmitted b y the spectrometer a t frequency v =
The quantity S, is the sensitivity factor of the monochromator/multiplier phototube combination; the true emission spectrum is calculated from the apparent emission spectrum by dividing it, point by point, by 8,. The simplest method of determining S,as a function of frequency in the visible region of the spectrum is to take multiplier phototube readings when the entrance slit of the spectrometer is illuminated by a tungsten lamp of known spectral distribution ( 7 ) . For wavenumbers greater than 2.5 p - 1 (400 mp), the intensity of light from a tungsten lamp falls off rapidly. Such a lamp is therefore not suitable for calibration a t the higher wavenumbers. Unfortunately, since discharge lamps with accurately known spectral distribution in the ultraviolet’
