for two reasons: a ) the Bates-Guggenheim convention has been assumed valid at a n ionic strength (ten times greater than originally intended; b) if used in cell 1, the liquid junction potential effect will be large. Neither of these have any bearing on the new method of testing. Accurate knowledge of the pH(S) values of the test solutions is not essential. Application of the New Method of Testing to Commercially Available pH Glass Electrodes. All the principal U.K.manufacturers kindly donated samples of their current ranges of pH responsive glass electrodes for testing by the new method. To these were added a few other electrodes available in the U.K. Duplicate electrodes were conditioned according to instructions given in the manufacturer’s literature. These referred to 0 . M HCl or distilled water for various specified times. In the absence of a recommendation, electrodes were conditioned in distilled water overnight. Results are given in Table V of tests of 15 commercial electrodes in the buffers B1 B6 which contain no Na+ ions, and in Table VI for the tests a t constant pH and increasing Na+ concentration. Also given in the tables are the manufacturers’ recommended range of use for their electrodes, but the type and source of the electrodes has been otherwise concealed by the allocation of code letters. Duplicate electrodes (denoted by a subscript) usually showed identical performance from the first time of use after the recommended conditioning treatment. In a few cases, one of a pair showed higher errors associated with drifting potentials, but the situation improved on further conditioning. Electrode Sp was not a new electrode, but had been in routine use for over a year. Electrodes recommended for the pH range 0-11 or 12 (e.g., N, S, U, V, X) often showed large potential-time variations a t the highest pH and Na+ concentrations. Even at pH 9.5 and pNa 0, electrodes N, P, and V show errors of about 0.1 pH and electrode X is even worse.
-
Of the electrodes recommended for the full pH range (014), Q and T are clearly the best, with K, L, M, R, W close behind. The electrode 0 recommended as “high alkaline” was no better than the full-range electrodes. It may be pointed out that some electrodes (0, Q, T) show apparently larger errors in solution B6 (Table V) than in the corresponding sodium ion-containing buffers (Table VI). One possible explanation for this is that the transfers in sodium-free solutions were done some weeks after the other tests, and it is well known that errors increase with the age of the electrode It should be recalled that B6 does contain a small amount of alkali metal ions (5 x 10-4M). The effect of carbon dioxide contamination also leads to an apparently increased error of the glass electrode, but this was not the reason here. We conclude that the new method of testing presented here, and appraised by studies on 30 commercial samples of the currently available U.K.range of glass electrodes, satisfactorily provides the analytical chemist with a simple means for evaluating glass electrode performance in alkaline buffers with and without sodium ions.
ACKNOWLEDGMENT This work was carried out with the encouragement of the British Standards Institution Committee LBC/16 concerned with Glass Electrodes and pH meters. We are indebted to the various past and present members of the committee for their helpful suggestions and in particular to its former Chairman, the late J. E. Prue for his advice. We acknowledge the support of the U.K.Manufacturers of glass electrodes with their gifts of glass electrodes.
RECEIVED for review December 31, 1973. Accepted February 19, 1974. We are grateful to the Instituta de Alta Cultura, Portugal, for granting financial support, and to the University of Lisbon for study leave, to one of us (MFGFCC).
Alternating Current Polarographic Determination of Uranium in Complex Minerals Characterized by Electron Probe Analysis A. M. Bond Department of Inorganic Chemistry, University of Melbourne, Parkville, 3052, Victoria, Australia
V. S. Biskupsky and D. A. Wark Department of Geology, University of Melbourne, Parkville, 3052, Victoria, Australia
Polarographic methods generally do not have the required specificity for undertaking the direct determination of a particular element in complex systems such as those encountered in geochemical analysis. In this work, it is shown that phase-selective high frequency ac polarography enables the determination of uranium in extremely complex minerals such as euxenite, samarskite, thorite, thorianite, and monazite and in glasses without the usually required separation procedures. The minerals are completely characterized by electron probe data, to show the wide range of other elements tolerable. The attempted determination of uranium by atomic absorption spectrometry, X-ray fluorescence, and
fission track analysis is also reported. The dlfflculties encountered in these methods because of the large matrix corrections required, or other reasons, demonstrate the usefulness of the proposed polarographic method. The specificity, not usually attributed to polarography, results from the ability of the high frequency ac method to discriminate against nonreversible electrode processes and species reduced at more positive potentials than uranium. However, high quality polarographic instrumentation is needed as the use of the conventional two-electrode cell arrangement can give rise to undesirable Instrumental artefacts, possibly explaining previously reported interferences.
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 11, SEPTEMBER 1974
1551
In general, the composition of uranium-bearing samples is complex (1-3). In processable ores, uranium is found with niobium, vanadium, thorium, zirconium, and a wide variety of other elements ( 3 ) . Consequently, most analytical methods for the determination of uranium require a preceding separation from other elements ( 3 , 4 ) and a wide range of separation methods has been proposed ( 3 ) . One of the most frequently used analytical methods for the routine and inexpensive determination of elements in geological samples is atomic absorption spectrometry. However, for uranium (5, 6) the sensitivity is poor, and the atomic absorption method is not well suited to the determination of this element in complex mixtures. Other spectroscopic methods can be used. However, in uranium minerals, matrix problems can lead to considerable difficulties (3,unless prior separations are carried out, or complex and often suspect correction factors are applied. Neutron activation or fission track analysis, electron probe, mass spectrometry, and other techniques can also be applied, although frequently with difficulty. However, these are highly specialized techniques and not generally available in all laboratories. Electrochemical techniques have been used frequently for the determination of uranium in a wide range of areas (3).In particular, polarographic methods are well suited to the determination of uranium (3, 4, 7, 8 ) . However, dc polarography is not normally sufficiently specific to determine particular elements in complex mixtures of the type encountered in geochemical analysis and, where polarographic methods have been proposed, the majority have been based on the dc form, with separation procedures included in the preliminary work up ( 7 , 9 , 1 0 ) . Modern polarographic techniques are far more specific and rapid than the dc form. Booman and Rein ( 3 ) ,for example, in discussing the choice of polarographic apparatus for the determination of uranium, note that fast-sweep oscillographic methods save the long scanning times necessary and give some discrimination against the interfering electrode processes, which are much slower than the reduction of uranyl ion. They also state that square wave and ac polarographic techniques increase the sensitivity by two or three orders of magnitude and increase the selectivity compared to conventional dc methods. However, despite these advances, little endeavor has been made to apply modern polarographic methodology directly to the determination of elements in extremely complex mixtures as has been pointed out recently by Maienthal ( 1 1 ) and where linear sweep, ac, pulse, and square-wave (4, 12-1 7 ) techniques have been ( 1 ) "Geochemistry and Mineralogy of Rare Elements and Genetic Types of Their Deposits," Vol. I, K. A. Vlasov, Ed., Acad. Sci. U.S.S.R., translated
from Russian, Israel Program for Scientific Translations, Jerusalem, 1966. (2) E. S.Dana, "A Textbook of Mineralogy," 4th ed.,revised by W. E. Ford, Wiley, New York. N.Y., 1955, p 851. (3) G. L. Booman and J. E. Rein, in "Treatise on Analytical Chemistry," Part (I, Section A , Vol. 9, I. M. Kolthoff and P. J. Elving, Ed., interscience, New York/London, 1962, pp 1-188. (4) "Proceedings of a Symposium on the Analytical Chemistry of Uranium and Thorium, Lucas Heights, Sydney, May 1970," T. M. Florence, Ed., Aust. At. Energy Comm. Rep., TM 552, August 1970. (5) M. D. Amos and J. B. Willis, Spectrochim. Acta, 22, 1325, 2128, (1966). (6) "Analytical Methods for Flame Spectroscopy," Varian Techtron, Australia, 1972. (7) M. Pinta, "Detection and Determination of Trace Elements," Dimod. Paris 1962, translated from French, Israel Program for Scientific Translations, Jerusalem, 1966. (8) B. Breyer and H. H. Bauer. "Alternating Current Polarography and Tensammetry," Interscience, New York/London 1963, pp 189-190 and references cited therein. (9) R. W. Martres and J. J. Burastero, Analyst(London),96, 579 (1971). (10) I. Hodara and I. Baloka, Anal. Chem., 43, 1213 (1971). ( 1 1) E. J. Maienthal, Anal. Chem., 45, 644 (1973). (12) T. M. Florence, Anal. Chim. Acta, 21, 418 (1959). (13) T. M . Florence and P. J. Shirvington, "Determination of Beryllium, Thori1552
applied in uranium analysis, separations have generally remained as an integral part of the determination. The reports on the square wave method (14-16), however, do give examples which demonstrate considerable specificity for uranium. For uranium, a highly specific, rapid, and direct polarographic method that could be used routinely in most laboratories would be highly desirable. In a recent article (It?), it was shown that the use of high frequency, short drop time, phase-selective ac polarography could be used to directly determine tin in geochemical samples on a time scale approaching that of atomic absorption spectrometry. Interlaboratory studies showed results to be in agreement with atomic absorption spectrometry, X-ray fluorescence, and colorimetric methods. In the present work, application of this variation of ac polarography for the direct determination of uranium in some extremely complex minerals is described. To assess the method, the minerals are completely characterized by electron probe data to reveal the extent and variety of foreign elements present in the samples. The attempted application of atomic absorption spectrometry, X-ray fluorescence, and fission track analysis on the same minerals, illustrates the difficulties associated with direct uranium determinations and demonstrates the value of the proposed polarographic method.
INSTRUMENTATION Polarographic. Alternating current polarograms were recorded with PAR Electrochemistry System Model 170 (Princeton Applied Research Corporation, Princeton, X.J.). A three-electrode system was used with tungsten wire as the auxiliary electrode and Ag/ AgCl (5M NaC1) as the reference electrode. Phase-selective detection was used and the in-phase component of the signal recorded. The amplitude of the alternating potential was 10 mV p-p, a t frequencies stated in the text and figures. Solutions were degassed with argon, although this step could be avoided with little or no loss of precision of measurement (19). Solutions were not thermostated (see below). Short controlled drop times between 0.16 and 0.32 sec were achieved with Metrohm Polarographie Stand E354. Atomic Absorption Spectrometry. Atomic absorption measurements were carried out using Techtron AA-100 and AA-4 spectrophotometers. The 3514.6-A line was used with air-acetylene and nitrous oxide-acetylene flames. Fission Track Analysis. A Zeiss Photomicroscope(II), with 80X epiplan objective was used to count fission tracks in transmitted light in both mica and "Lexan" plastic (20). Samples were irradiated in the Australian Atomic Energy Commission Reactors. The uranium standard was glass "1497" containing 0.84% natural uranium (21). Electron Probe. A JEOL JXA-5A microprobe was used with PDP-8 computer control of three X-ray spectrometers. Corrections for dead time, absorption, atomic number, and fluorescence were applied by the Program of Mason, Frost, and Reed (22). T h e uranium standard was the same metal as used for the polarography. The uranium MP line was measured by PET crystal and argonmethane counter. The measurement of Rare Earths and other elements is described elsewhere (23). X-Ray Fluorescence. A Siemens Sequential X-Ray Spectrometer Model SRS-1 was used to measure the L a line of uranium by urn. and Uranium in Suiphuric-Phosphoric Acid Mixtures," Aust; At. Energy Comm. Rep. TM 153, Sydney, September 1962. (14) G. W. C. Milner and J. H. Nunn, Anal. Chim. Acta, 21, (1959). (15) D. J. Ferret and G. W. C. Milner. Analyst, (London), 80, 132 (1955). (16) 0.Gurtler and Chu-Xuan-Anh, Mikrochim. Acta, 941 (1970). (17) G. W. C. Milner, J. D. Wilson, G. A. Barnett and A. A. Smales, J. Elecfroanal. Chem., 2, 25 (1961). (18) A. M. Bond, Anal. Chem., 45, 2026 (1973). (19) A . M . Bond, Talanta,20, 1139(1973). (20) J. D. Kleeman and J. F. Lovering, At. Energy Aust., 10, 3 (1967). (21) R. L. Fleischer, Geochim. Cosmochim. Acta, 32, 989 (1968). (22) P. K. Mason, M. T. Frost, and S.J. E. Read, B.M.-i.C.-N.P.L. Computer Programs for Calculating Corrections in Quantitative X-Ray Microanalysis, Report 2, Division of Inorganic and Metallic Structure, National Physical Laboratory, U.K., April 1969. (23) D. A . Wark, J. F. Lovering, A. F. Reid and A. El Goresy, in "Lunar Zirconolite: A late-stage (mesostasis) phase in lunar igneous rocks," in preparation.
ANALYTICAL CHEMISTRY, VOL. 46, NO. 11, SEPTEMBER 1 9 7 4
LiF crystal and scintillation counter. T h e samples were 23-mm diameter, pressed powder pellets made from 0.5 gram of mineral diluted with 1.5 grams of alumina t o reduce matrix effects. A chromium X-ray tube (at 55 kV and 30 mA current) was used t o avoid the interferences the usual gold tube would produce with t h e uranium X-ray lines. T h e uranium standard was the same U308 as used for polarography. Chemicals. All chemicals used were of reagent grade purity. Further specific details are given a t appropriate places in the text. Procedures. Sample Dissolution. T h e minerals used are extremely complex and techniques for their dissolution required careful consideration. All of the minerals were subjected to the procedures described below. Sodium Peroxide Fusion: 50-100 mg of fine grained sample (100 mesh) was mixed with 0.5 gram of Na202 in a zirconium crucible. A ten-minute fusion with the crucible covered was then carried out. T h e temperature was regulated during the fusion so that the bottom of the crucible was dark red. After cooling, 10 ml of water was added and the solution was then boiled for several minutes to decompose excess NazOz. After cooling, HzS04 was added to acidify the solution ( p H 2 to 3). Finally, the solution was transferred quantitatively to a 25-1111 standard flask. Lithium Fluoride-Boric Acid Fusion: 50-100 mg of the sample was fused with 0.2 gram boric acid and 0.3 gram LiF in a platinum crucible for 5 minutes. After cooling, 1 ml of concd H2S04 was added and t h e mixture heated until copious fumes of SO3 appeared. After this, the mixture was again cooled and a small amount of water added. T h e solution was then boiled to complete the dissolution. Finally, the solution was transferred to a 50-ml standard flask and made up to volume. This method is based on that, by McLaughlin and Biskupsky (24). Agreement of results for the two fusion methods was taken as evidence that complete sample dissolution for uranium was being achieved. On some of the samples, a very small quantity of precipitate remained which when tested by X-ray fluorescence was uranium-free. Polarographic analysis, unlike atomic absorption or spectrophotometric methods can be carried out directly without filtration or settling of any solid materials. Standards for Polarography. Four uranium standards were investigated, uiz., uranium metal, uranium trioxide, uranyl nitrate ( U 0 2 ( N 0 3 ) 2 . 6 H 2 0 )and , uranyl acetate (UOz(CH3COO)y2HzO). T h e uranium metal after preliminary treatment with nitric acid and drying with acetone to remove oxide coating was dissolved in 10 ml of 10M H N 0 3 . Five ml of 1:l H2S04 was then added and the solution evaporated until so:,vapors appeared. Uranium trioxide was heated to constant weight and uranium standards were prepared by weighing as U308. Based on triplicate independent preparations of nominally 10-"M uranium standards from the four uranium sources and comparison of their polarograms in 0.1M H2S04, the uranium metal and U:~OSmethods were accepted as being satisfactory standards. Both gave reproducible solutions of identical concentration within the limit of experimental error ( f l % ) . Uranyl nitrate, while giving results close to those for the metal and oxide, showed lower reproducibility indicating slightly variable content of water of crystallization. Uranyl acetate gave somewhat lower uranium concentrations, indicating perhaps the presence of more water than the two waters of crystallization nominally present in the compound. For t h e determination of uranium in minerals, bulk standards containing 10 mg/ml, 1 mg/ml, and 0.1 mg/ml were prepared in the medium provided by the fusion. These solutions were used to prepare calibration curves and in the determinations by t h e method of standard additions.
R E S U L T S A N D DISCUSSION T h e o r y f o r Polarographic Determination of U r a n i um. Alternating current polarographic methods are undoubtedly best suited to reversible electrode processes (25). In mineral acid media, uranium gives several polarographic waves (3, 8). The most positive wave is generally associated with the reversible charge transfer step U W ) + e z s U(V) (1) and is suitable for use by ac polarography. Dissolution techniques used in geochemical analysis alMcLaughlin and V. S. Biskupsky, Anal. Chim. Acta, 32, 165 (1965). (25) A. M. Bond, Anal. Chern., 44, 315 (1972). (24) R. J. W.
most invariably lead to the direct preparation of acidic solutions of mineral acids as is the case in this work. Such media are therefore already suitable for the use of the U(VI)/U( V) couple and additional supporting electrolytes or other reagents are not required in principle. However, detailed analysis of the electrode process (3, 8) shows it to be more complex than Equation 1 would indicate and U(V) is prone to disproportionation. That is, the electrode process is more generally represented by Equation 2.
In acidic media uranium(V1) exists as UO*"+, and, in the absence of hydrolysis, the charge transfer step can be written
UO~~ + +e z=uoZt
(3)
The disproportionation rate ( 3 )of U02+ is second order with respect to uranium(V) concentration and first order with respect to hydrogen ion concentration (activity). +
- d[Uoz dt
+ z = constant [Ht][U02 ]
(4)
Thus, with ac polarography, the peak height of the U(VI)/ U(V) wave is a function of pH, ionic strength, and other variables unless the disproportionation step can be suppressed until negligible. Provided the charge transfer step is rapid, the use of high frequency ac polarography, instead of the more usually used low frequency range, minimizes the influence of disproportionation as the time scale available for the follow up reaction is decreased. Similarly, the use of short drop time decreases the dc time scale and discriminates against disproportionation. Thus, short drop time, high frequency ac polarography should be considerably less influenced by disproportionation than conventional dc polarographic techniques. However, in geochemical analysis, the solution can literally contain almost any element in the periodic table, in addition to the uranium being determined. Any element reacting with U(1V) for example could catalyze the disproportionation step and interfere with the determination. Furthermore, it is difficult to rigorously control the ionic strength, pH, and other conditions in geochemical analysis, and a wide range of potential interference emanating from uncertainties in the extent of disproportionation can occur. T o minimize the above possibilities, the method of standard additions was employed in conjunction with the high frequency, short controlled drop time ac method as a further check for interference. Frequencies in the range 500 to 1000 Hz were used and drop times of 0.16 to 0.32 sec. Polarographic Method f o r Determination of U r a n i um. I t was established over a wide range of conditions of pH, ionic strength, and frequency, that linear, but not exactly identical, calibration curves (peak height us concentration) could be obtained for the uranium concentration range 10-3M to 10-6M. For actual analysis, the uranium standards containing 10 mg/ml, 1 mg/ml, and 0.1 mg/ml were added to three micro burets. Exactly 26 ml of the appropriate blank fusion mixture was added to the polarographic cell, and appropriate aliquots from one of the uranium standards were added. Figure 1 shows the addition of aliquots of the 10 mg/ml uranium standard to both the NazO2 and LiF/H3B03 acid fusion mixtures and the use of high frequency, rapid ac polarography. Despite the consid-
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 11, SEPTEMBER 1974
1553
Na202 FUSION
L I F / H3B03 FUSION
Samarskite i N T i 1Wmg Na202 fusion 25m
Drop Time = 0 2Lsec
NoturOI Drop 1178
Scan Rote
frequency
AE
=
i
20rrvlsec
i
qOOhz
A E :10rnV
l0nV
FreqLency = 1090 hz
U
0.8ml U Std
VOLT vs AgiAgCl - ;2
CG
-0"
Volt vs Ag/AgCI
EUXENITE 1COrrg. lua2C; f u s i c r , 2% Ncturcl 3 r o p Time Frequency =qOChz
& E = lOmv
22
-
~
~~
co
.L 2
-_------_.
V3LT vs AgiAgC!
U Std
-0 2
00
-0 L
VOLT
vs
-c
2
-2 L
Ag/AgCI
Figure 1. Addition of a 10 mg/ml U standard to 25 ml of the peroxide and lithium fluoride-boric acid fusion mixtures
erable differences between the electrolytes, almost linear and identical calibration curves are obtained. A t each concentration, the peak potential is identical and, importantly, the half-width is invariant a t (93 f 3) mV, which is the value expected for a reversible one-electron reduction (25). Checks on interference are therefore readily built into the ac polarographic method as departure of peak potential or, even more importantly, half-width are indicative of interference. From calibration curves prepared in the above manner, approximate concentrations of uranium could be determined. This result was taken as a guide to the choice of standard to be used in the final determination by the method of standard additions. Figure 2 shows examples of natural drop time ac polarograms of some of the minerals used in this work. In each case, the uranium wave is extremely well defined, with a peak potential of close to -0.20 V us. AgIAgC1. Importantly, the half-width as for the standards, is also found to be (93 f 3) mV in each case. The presence of lead is also indicated. As radiogenic lead usually occurs with uranium, this is expected, and the simultaneous determination of U and P b would be possible in many samples. Figure 2 shows a high frequency short drop time ac polaro1554
00
-2 i
VOLT vs
-3 1
AglAgCl
Figure 2. AC polarograms of some of the minerals used in this work
0 2ml
00
-ii
Pb
gram of one of the samples. As for the equivalent polarograms in Figure 2, the uranium wave is exceedingly well defined. Figure 3 shows the method of standard additions as applied to the determination of uranium in this sample. The peak height of the uranium wave increases in height; but the peak potential and half-width remain unaltered. Under conditions where a linear standard additions plot is obtained and the peak potential remains constant and the half-width is (93 f 3) mV, the uranium determination is concluded to be interference-free and quantitative. Such results were achieved with the high frequency short controlled drop times on all samples examined. Results are tabulated in Table I for the minerals examined. In view of the use of the method of standard additions, the need for thermostating of solutions was obviated, as ambient temperatures over the times required for the uranium determination did not vary by more than f 0 . 2 "C. Verification of Method via Synthetic Mixtures. Based on complete electron probe analyses (see below) appropriate synthetic mixtures containing 0.1%, 1%,and 10% uranium and the major elements detected were prepared. Other elements were added as CeOn, ThOz, ZrOn, TiO2, Fe203, Ca3(PO4)2, MgO, TazOj, Nbz05, Si02, and A1203. None of these elements was found to interfere a t the 5% level, when using the method of standard additions for determination of uranium. This indicates that the method has a high degree of specificity for uranium, and a large number of elements over wide concentration ranges can be tolerated. Indeed for the samples investigated, no interference was encountered within the limit of experimental error since the answers returned were within 2% (relative) of the accurately known weighed uranium percentage.
ANALYTICAL C H E M I S T R Y , VOL. 46, NO. 1 1 , SEPTEMBER 1974
0.3ml
THORIANITE (CEYLON)
0 2Lsec
Drop Time
U Std
0.2ml
50W. L I F I H ~ B OFUSION, ~ 50ml
U
0.1ml
Std
U
Std
U
U
Scan Rote = 20rnv/sec A E = lOmv Frequency = 500 hz
U
I
-02
00
t
I
-Q 2
00
-0 L
I
I
-04
00
-0 2
vs
Ag/AgCI
Volt
-04
-0 2
00
-04
Figure 3. Determination of uranium by method of standard additions: U standard = 10 mg/ml ~~
Table I. Polarographic and E l e c t r o nProbe D e t e r m i n a t i o n s of Uranium in Some Complex M i n e r a l s Polarography Average
u, %
Mineral”
Euxenite Fergusonite Monazite Samarskite “A” Samarskite “B” Samarskite “C” Thorianite “A” Thorianite “€3” Thorite
Electron probe V us. Ag/AgCl
Average U, %
0.205 0.205 0.208 0,202 0.205 0.205 0.195 0.200 0.203
7.48 3.99 0.30 14.61 13.31 6.38 25.28 8.43 9.19
-E,b
Range,
7.32 4.00 0.30 14 . O 12.6 6.30 27 . O 8.05 10.2
5%
7.20-7.44 3.96-4.04 0.28-0.32 13.4-14.6 12.1-13.5 6.00-6.60 26.0-28.0 7.60-8.60 9.60-10.6
Range,
L%,
7.40-7.56 3.92-4.05 0.295-0.309 14.33-15.02 12.28-14.49 6.22-6.48 25.15-25.41 7.75-9.12 8.88-9 .59
“ Euxenite, Arendal, Norway. Fergusonite, Norway. Monazite, Unknown Origin. Samarskite “A,” Hart’s Range, South Australia. Samarskite “B,” Last Hope Area, Hart’s Range. South Australia. Samarskite “C,” Lone Pine Area, Northern Territory, Australia. Thorianite “A,” Ceylon. Thorianite “B,” Andranondambo, Madagascar. Thorite, Langesund Fjord, Norway. All half-widths (93 i- 3) mV.
*
freqliency = 553 bz
A E = l5mv
,
, 2
3;
~
-7i
-- -
-c
VOLT vs AgiAgCl
1
~
.~ -3E
--------
‘3
- c c -01 -02 -03 -c2
7 -,*
VOLT vs Ag/AgCI
Figure 4. Interference encountered with vanadium in dc polarography is not found with the ac technique. Concentration of uranium(V1) = 10-4M. Concentration of vanadium(V)= lO-*M. LiF/H3B03 fusion media N e e d f o r High Quality Instrumentation t o Avoid Interference. The use of high frequency ac polarography
with phase-selective detection coupled with the determination of uranium complex solutions requires the use of high quality instrumentation to avoid interference arising from the presence of high concentrations of more positively reduced species (26).Vanadium(V) has been cited frequently (26) A . M. Bond and J. H. Canterford, Anal. Chern., 43, 228 (1971)
(31
3 electrode
h
-00 -31 - c 2 -33
V o l t s vs A g i A g C
Figure 5. Comparison of 2- and 3-electrode ac polarographyfor uranium with natural drop time and in the presence of vanadium(V). [U(Vl)], [VCV)],and other conditions as for Figure 4
as a potentially interfering species in the polarographic determination of uranium. Figure 4a shows a dc polarogram of a high concentration of V(V). The large dc current a t the potential where the uranium wave should occur completely masks the uranium wave. Figure 46 shows the ac polarogram is, however, still extremely well defined. Vanadium(Vj is irreversibly reduced in acid media and is severely discriminated against by high frequency ac polarography. However, the dc electrode process still occurs and can influence the electrochemistry via the large dc currents
ANALYTICAL CHEMISTRY, VOL. 46, NO. 11, SEPTEMBER 1974
1555
frequency = 500 hz AE = 10rnv Drop Time = 0 21 sec
1 0 1 3 electroae
PHOSPHATE GLASS Natural Drop Time AE =10mv
a
1
Frequency = 900 hz No202 fusion, 56rng. 25ni
t oc
-c12
-02
00
-04
V o l t s vs AgiAgCl
Figure 6. Comparison of 2- and 3-electrode ac polarography for uranium with short controlled drop time and in the presence of vanadium(\/). [U(W)], [V(V)] and other conditions as for Figure 4 00
rngiml uron urn stcraora
02 Present
02 Removed Orap Time :O 215ec
AE
:l 0 m v
Frequency = 5 0 0 h r YonRote i 2Omvlsec
-0 2 -c L VOLT vs AglAgCl
Flgure 8. Determination of uranium in a phosphate glass
I
flowing through the cell or via catalysis of the disproportionation step from interaction of products of the V(V) electrode process with the uranium reduction. Figure 5 shows the phase-selective two-electrode and three-electrode natural drop time ac polarograms of uranium in the presence of high concentrations of V(V). Uncompensated resistance terms enhanced by the large dc currents arising from the V(V) electrode process cause considerable apparent phase-angle shifts and invalidate the theory behind the use of in-phase measurements. With short controlled drop time techniques iR losses are reduced (27, 28). However, three-electrode instrumentation is essential as can be seen from Figure 6. Instrumental artifacts of this kind readily explain why polarography has had restricted use in complex solutions as it is only recently that high quality 3-electrode instruments have been in wide use. Removal of Oxygen. Figure 7 shows an ac polarogram, with and without removal of oxygen. A small degree of oxygen dependence of the electrode process is found. However, if the method of standard additions or calibration curves in the presence of oxygen is used, uncertainties are less than 5% and acceptable in many determinations of uranium. For high precision work ( f 2 % ) , removal of oxygen is still recommended. Determination of Uranium in Glasses. The determination of uranium in several glasses was undertaken, using the same procedures as for the minerals. Figure 8 shows an ac polarogram of one of the samples. Extension of the above methods to this analytical area is readily achieved.
The time taken to record a rapid ac polarogram is about 15 sec. Employing the method of standard additions required 4 to 5 recordings of an ac polarogram. Consequently, approximately two minutes of actual polarographic recording time per sample are required which is still significantly less than the time required to record a single conventional dc polarogram. The use of standard additions, of course, substantially increases the time for each determination. For semi-quantitative work, this can be avoided. However, the additional safeguards or checks on interference built into the standard addition method strongly recommend this technique. The limit of detection of the rapid ac polarographic method is approximately 10-6M for uranium standards. For complex solutions of the type encountered in geochemical analysis, the practical lower limit was found to be about 5 X 10-6M. The reproducibility of any one polarogram is f l %or better. However, cumulative errors introduced by sample dissolution and other procedures generally gave an overall reproducibility of f 5 % , based on triplicate analyses. Electron Probe Data. Many of the mineral samples proved to be slightly inhomogeneous. Polarographic data were recorded on samples crushed to millimeter size and hand picked to minimize any imperfect areas before powdering prior to sample dissolution. Determinations of uranium by ac polarography, therefore, represent average values for the samples selected. The electron probe more easily avoided altered areas on the mounted, polished mineral grains. Also, though a moving broad-beam technique was used, the relatively small area sampled would be less representative than the powder used for polarography. For these reasons, complete agreement between probe and polarography uranium values is not to be expected. However, these data are also most useful for assessing the elements present in the solutions for polarographic determination and, hence, the specificity of the polarographic method. Since a t least twenty other elements had to be analyzed to make matrix corrections for uranium, the use of the electron probe in this instance was very time-consuming and expensive and not competitive with ac polarography. For simpler systems, automated electrop probe analysis can be very efficient (29). Table I1 shows complete electron probe analyses and Table I compares pr.obe and polarographic figures for uranium. Within the limitations stated above, agreement of
(27) A. M. Bond and D. R. Canterford, Anal. Chem., 44, 1803 (1972) (28) A. M. Bond, Talanta, in press.
(29) B. L. Gulson. and J. F. Lovering, Geochim. Cosmochim. Acta, 32, 119 (1968).
6 50pA
I
00
-02
-31
00
-02
-0L
VOLT vs AglAgCi
Figure 7. AC polarography of uranium in the presence and absence of oxygen
1556
ANALYTICAL CHEMISTRY, VOL. 46, NO. 11, SEPTEMBER 1974
Table 11. Average Electron Probe Analyses in Weight Percent of Oxides Samplea Euxenite (2)
*
Oxide
NanO
0.09
< o .02
MgO
CaO MnO FeOc PbO A1203 Y,Os
La20J Ce.OJ PrBO:,
Nd.03 Sm20a EusOa Gd,Oy Tb2Oy DyzOs H0.03
ErzOj Tm?O:{ Y b2Oa Lu20a SiO, 'NO, %rOs
'rho:! UO? P20, Nb?Oj d
Total
27.59 1.21
. . .".
... ...
100.05
(1)
...