pared to the other noise terms in Equation 6 and SO the optimum slit width will then be more accurately expressed by Equation 8 than by Equation 9. If,the electrometer noise is the major source of noise, then iVm will be more simply given by
N, c 2
-
53.5Av~B(2’) ( Af) *”Ai, pLXo2g,At6 [ 7TyWH ( A / F 2 ) I 0]
(10)
For any given atomic absorption spectral line of a certain atom and for any given source-flame cell-monochromator-detector-electrometer combination, the limit of detectability should be essentially constant. An increase in slit width. U’,or slit height, H , or intensity, I ” , results in a proportional increase in the intensity of
radiation reaching the detector. This results in a proportional increase in the electrometer current setting for fullscale deflection and a slightly less than proportional increase in electrometer noise-Le., the electrometer noise (1, 2) will increase linearly with the anode current of the electrometer tube raised to some power which will probably be between IjP and 1. Therefore, an increase in slit width or height or an increase in Io with no resultant change in emission line half-width should give little change in the values of N , and
wo.
Whatever the situation, the analyst must a t least be aware of the electrometer noise. I n atomic emission flame spectrometery, the electrometer noise will rarely be the significant noise, whereas in atomic absorption flame
spectrometery, the total noise signal will in many instances be primarily a result of electrometer noise. LITERATURE CITED
( 1 ) Cannon, C. G.,
“Electronics for Spectroscopists,” Interscience, New York, 1960. ( 2 ) “RCA Radiation Designer’s Handbook,” fourth ed., F. Langford-Smith, ed., p . 940, Amalgamated 1\ ireless Valve Co. Pty., Ltd., Sydney, Australia, 1960. ( 3 ) Williams, F. C., J . Inst. Elec. Engrs. (London) 82, 561 (1938). ( 4 ) Winefordner, J. D., Vickers, T. J., ANAL.CHEM.36, 1939 (1964). ( 5 ) Ibid., p. 1947. J. D. WINEFORDNER CLAUDE VEILLON
Department of Chemistry University of Florida Gainesville, Fla.
Spectrophotometric Determination of Iron, Nickel, Copper, and Cobalt in Tungsten and Tungsten Alloys SIR: Iron, nickel, and copper in tungsten and tungsten alloys have previously been determined spectrophotometrically by somewhat lengthy extraction procedures using bathophenanthroline (S), diethyldithiocarbamate ( 5 ) , and neocuproine (a), respectively, as the colorimetric reagents. No method has been proposed for the spectrophotometric determination of cobalt in tungsten or tungsten alloys. In the present work iron, nickel, copper, and cobalt in tungsten and tungsten alloys are determined rapidly by spectrophotometric procedures by taking aliquots of a hydrofluoric-nitric acid solution that, has been buffered to a pH of approximately 6 by the addition of ammonium tartrate and sodium borate. KO subsequent adjustment of pH is necessary for the spectrophotometric determination of the iron, copper, or cobalt. EXPERIMENTAL
Reagents. Sodium Borate Solution (4.5y0). Dissolve 90 grams of sodium borate (Na2B4O?, anhydrous fused powder) in about 1800 ml. of hot vrater, cool, and dilute to 2 liters. Blank Solution. Transfer 5 ml. of hydrofluoric acid and 4 ml. of nitric acid to a platinum dish and evaporate to a volume of about 4 to 7 ml. on the hot plate (the purpose of the heating is to attain similar conditions as in the actual method). Add about 30 ml. of water, wash into a 250-ml. volumetric flask containing 50 ml. of ammonium tartrate solution (20y0),add 90 ml. of sodium borate solution (4.570), cool, and dilute to the mark.
Preparation of Calibration Curves. IRON. Transfer 50 ml. of blank solution to six 100-ml. volumetric flasks. Use one solution as a blank. Add 0.5-, 1.0-, 2.0-, 3.0-, and 3.5-ml. portions of standard iron solution (1 ml. = 0.10 mg. of Fe). Dilute to about 75 ml. and add 1 ml. of hydroxylamine hydrochloride solution ( l O ~ o ) and 5 ml. of o-phenanthroline solution (0.2%). Dilute to about 90 ml., heat in a water bath a t 60” to 70” C. for 30 minutes, cool, and dilute to the mark. Measure the transmittance a t 510 mp with a spectrophotometer (1-cm. cell) that has. been set to 100% transmittance with the reagent blank. Plot milligrams of iron against per cent transmittance. NICKEL(CURVE No. 1). Transfer 50 ml. of blank solution to six 100-ml. volumetric flasks. Use one solution as a blank. Add 1.0-, 3.0-, 4.0-, 5.0. , and 6.0-ml. portions of standard nickel solution (1 ml. = 0.10 mg. of Ni). Add 5 ml. of bromine water. Add ammonium hydroxide until the bromine color is bleached, and then add 5 ml. of excess ammonium hydroxide. -4dd 5 ml. of alcoholic dimethylglyoxime solution (lye). Add the bromine water, ammonium hydroxide,. and dimethylglyoxime within a period of 2 minutes. Dilute to the mark and within 15 f 2 minutes measure the transmittance at 540 mp with a spectrophotometer that hm been set to 100yotransmittance with the reagent blank. Plot milligrams of nickel against per cent transmittance. NICKEL (CURVENo. 2). Proceed as for curve No. 1 up to and including the addition of the bromine water. .4dd 10 ml. of ammonium hydroxide and 5 ml. of alcoholic dimethylglyoxime solu-
tion (lye), allow to stand for 1 minute, and add 10 ml. of sodium hydroxide solution (25%). Dilute to 100 ml., allow to stand for 5 minutes or more, and measure the transmittance as for curve No. 1. COPPER. Transfer 50-ml. portions of blank solution to five 100-ml. separatory funnels that have been rinsed with 95% ethyl alcohol to remove water from the stems. Add 2.0-, 5.0-, 7.5-, and 10.0-ml. portions of standard copper solution (1 ml. = 0.01 mg. of Cu). Add 5 ml. of hydroxylamine hydrochloride solution (lo%), 10 ml. of neocuproine solution (O.lyoin ethanol), and 10 ml. of chloroform. Shake for 30 seconds, allow the layers to separate, and drain off the chloroform layers into 25-ml. volumetric flasks containing 3 ml. of 95% ethyl alcohol. Repeat the extraction with 5 ml. of chloroform. Dilute to the mark with 95% ethyl alcohol and measure the transmittance a t 455 mp with a spectrophotometer that has been set to lOOyo transmittance with the reagent blank. Plot milligrams of copper against per cent transmittance. COBALT. Transfer 10-ml. portions of blank solution to seven 150-ml. beakers. Use one solution as a blank. Add 1.0-, 2.0-, 3.0-, 4.0-, 5.0-, and 6.0-ml. portions ’ of standard cobalt solution (1 ml. = 0.01 ml. of Co). Add 2.0 ml. of nitroso-R salt solution (l’%), cover with watch glasses, heat to boiling, and boil moderately for 60 seconds. Remove from the hot plate and add 2.0 ml. of hydrochloric acid. Return to the hot plate and heat just to boiling. Cool to room temperature, wash into 25-ml. volumetric flasks, and dilute to the mark. Measure the transmittance a t 510 mp with a spectrophotometer VOL. 37, N O . 3, MARCH 1965
417
Table 1.
Results (%)for Iron, Nickel, Copper, and Cobalt in Samples of Tungsten and Tungsten Alloys
Sample 1 (tungsten metal)
Fe 0010 0015 0010 0015 0013 5 08 5 18 5 18
0 0 0 0 Av. 0
2a
Av. 5 15 1.08 1.00 1.08 Av. 1.05 5.15 5.01 5.15 Av. 5.10
Ni
cu
0 004 0 004 0 004
0 0000 0 0000 0 0000
0 004 4 91 4 91 4 78 4 78 4 85 6.75 6.75 6.92 6.75 6.79 4.63 4.63 4.51 4.51 4.57
0 0 0 0
0000 0000 0000 0000
0 0000
2.20 2.00 2.10 2.10 2.10 0.0035 0.0035 0.0040 0.0040 0.0038
Av .
Av.
co 0.0002 0.0002 0.0000
0 0 0 0 0 0
0001 51 51 53 51
52 0.010 0.011 0.011 0.010 0.011 0.53 0.53 0.54 0.54 0.54 11.64 11.92 11.88 11.84 11.82 11.76 11.68 11.80 11.76 11.75
Contains (%) 5.13 Fe (titr.); 4.83 Ni (grav.); 0.52 Co (potent.). Contains (yo)1.03 Fe (titr.); 6.76 Ni (grav.); 2.16 Cu (elect.). c Contains (70) 5.13 Fe (titr.); 4.55 Ni (grav.); 0.54 Co (potent.). d Contains (%) 11.77 Co (potent.). Contains (Yo) 11.68 Co (potent.).
0.5 mg. of nickel, and 0.06 to 0.10 mg. of copper; dilute to about 50 ml.; and proceed rn above. Use curve KO. 2 for nickel. For cobalt, pipet an aliquot containing preferably 0.03 to 0.06 mg. of cobalt (the aliquot should be 10 ml. or less), add 10 ml. of blank solution to act as a buffer, and proceed as above. If a small size aliquot is used for the cobalt determination, the tungsten remains in solution and the filtration is unnecessary. DISCUSSION A N D RESULTS
I n the spectrophotometric method for iron the maximum amounts of nickel, copper, and cobalt that may be present in the aliquot without causing interference are 0.8 mg. of nickel, 1.6 mg. of copper, and 0.8 mg. of cobalt. More than these amounts cause low results. The usual technique of determining nickel spectrophotometrically by adding bromine water, ammonium hydroxide, and dimethylglyoxime ( 1 ) gave satisfactory results for small amounts of nickel but variable results for large amounts of nickel where the stability of the color is all-important. Satisfactory results were obtained for larger amounts of nickel by adding sodium hydroxide to stabilize the color ( 4 ) ;
a b
6
Table It.
Recovery of Iron, Nickel, Copper, and Cobalt
Added, mg.
Recovered, mg. Fe
that has been set to 1 0 0 ~ transmittance o with the reagent blank. Plot milligrams of cobalt against per cent transmittance. If desired, 10 ml. of 20% ammonium tartrate solution (diluted to about 50 ml.) may be used in place of the blank solution in preparing the curves for the iron, nickel, and copper. The blank solution must be used in preparing the curve for cobalt. Procedures. For the determination of less t h a n 0.09% iron, o.1570 nickel, 0.03% copper, a n d o.O4y0 cobalt weigh a 2-gram sample into a platinum dish, add 5 ml. of hydrofluoric acid and 4 ml. of nitric acid, and warm to dissolve (away from direct heat). Also, heat a blank in t h e same manner. If prolonged heating is necessary add more hydrofluoric and nitric acids. The final volume should be about 4 to 7 ml. Add 30 ml. of water and wash into a 250-ml. volumetric flask containing 50 ml. of ammonium tartrate solution (20%). Add 90 ml. of sodium borate solution (4.5y0), dilute to about 220 ml., cool to room temperature, and dilute to the mark. IRON. Pipet a 50-ml. aliquot into a 100-ml. volumetric flask. Add 1 ml. of hydroxylamine hydrochloride solution (10%) and proceed as in the preparation o f ' the calibration curve. Calculate the per cent iron by referring to the calibration curve. 418
0
ANALYTICAL CHEMISTRY
NICKEL.Pipet a 50-ml. aliquot into
a 100-ml. volumetric flask. Add 5 ml.
of bromine water and proceed as in the preparation of calibration curve No. 1. Calculate the per cent nickel by referring to the calibration curve. COPPER. Pipet a 50-ml. aliquot into a 100-ml. separatory funnel. Add 5 ml. of hydroxylamine hydrochloride solution (loyo) and proceed as in the preparation of the calibration curve. Calculate the per cent copper by referring to the calibration curve. CoBAur. Pipet a 20-ml. aliquot into a 150-ml. beaker. Add 2.0 ml. of nitroso-R salt solution (l%), develop the color, and dilute to 25 ml. in a volumetric flask as in the preparation of the calibration curve. Filter a portion of the solution through a dry Whatman No. 42 filter paper into a 50ml. beaker. discarding the first few milliliters. Measure the transmittance at 510 mp on a portion of the filtrate and calculate the per cent cobalt by referring to the calibration curve. For tungsten-iron, tungsten-nickel, tungsten-copper, and tungsten-cobalt alloys proceed as above but use a 1-gram sample and dilute to 500 ml. in a volumetric flask. Pipet a 25-ml. aliquot into a 250-ml. volumetric flask (1-liter volumetric flask for cobalt higher than 8%) and dilute to the mark. For the determination of iron, nickel, and copper, pipet aliquots containing preferably 0.15 to 0.3 mg. of iron, 0.3 to
0.010" 0.050 0.100 0.200
0.012 0.053 0.098 0.205 0,302 0.300 Ni (",OH Method) 0,010" 0,012 0,100 0.102 0,200 0.200 0,400 0.397 0.590 0.600 Ni (NHdOH-NAOH Method) 0 100a1b 0 097 0.200 0.200 0,400 0,406 0.600 0.598
cu 0.005a 0.010 0.050 0.070 0.100
0.004 0.008 0,051
0,070 0.097
co 0 . 005c
0.005 0.010 0.011 0.030 0.031 0.040 0,040 0.060 0.061 4 Added to 50-ml. aliquots of pure tungsten solution. b This method designed for larger amounts of nickel; no data given for less than 0.1 mg. of nickel. c Added to 20-ml. aliquots of pure tungst,en solution.
however, erratic results were obtained for trace amounts using this technique. Therefore, two procedures for the determination of nickel are recommended; the ammonium hydroxide method for low percentages and the ammonium hydroxide-sodium hydroxide method for higher percentages. I n the spectrophotometric determination of nickel by the ammonium hydroxide method the maximum amounts of iron, copper, and cobalt that may be present in the aliquot are 20 mg. of iron, 0.2 mg. of copper, and 1.5 mg. of cobalt. I n the ammonium hydroxide-sodium hydroxide method the maximum permissible amounts are 15 mg. of iron, 0.2 mg. of copper, and 1.5 mg. of cobalt. More than these amounts cause erratic results. There are no interferences with the neocuproine method for copper ( 2 ) .
I n the cobalt method, on adding the hydrochloric acid to large size aliquots and boiling to destroy the heavy metal complexes of nitroso-R salt, some tungsten precipitates and must be filtered off. Because color is developed prior to the precipitation of the tungsten, this precipitation causes no error. The maximum permissible amounts of iron, nickel, or copper that may be present in the aliquot in which the cobalt color is developed are 15 mg. of iron, 2 mg. of nickel. and 3 mg. of copper. More than these amounts cause low results. The results obtained for iron, nickel, copper, and cobalt in samples of tungsten and tungsten alloys are shown in Table I. The results by the spectrophotometric methods checked the results by classical methods where the range permitted the use of both techniques. The recoveries obtained on
adding standard iron, nickel, copper, and cobalt solutions to tungsten metal and carrying the samples through the procedures were satisfactory (Table
11). LITERATURE CITED
(1) Am. Soc. Testing Materials, “1964 Book of ASTM Standards, Part 32, Chemical Analysis of Metals,” p .
434, Philadelphia, Pa.
( 2 ) Gahler, A. R., ANAL.CHEM.26, 577
(1954).
(3) Gahler, A. R., Hamner, R. M., Shubert,, R. C., Ibid., 33, 1937 (1961). (4) Makepeace, G. R., Craft, C. H.,
IND. ENG. CHEM.,ANAL.ED. 16, 375 (1944). ( 5 ) Rohrer, K. L., ANAL.CHEM.27, 1200 (1955). GEORGE NORWITZ HERMAN GORDON Pitman-Dunn Laboratories Frankford Arsenal Philadelphia 37, Pa.
Rapid Removal of Transuranium Elements from Aqueous Solutions Prior to the Determination of Total Lanthanide Fission Products SIR: The radiochemical determination of total lanthanide fission products in solutions containing high levels of the very radiotoxic transuranium alpha emitters presents a difficult problem to the radiochemist. It is desirable to remove the bulk of the alpha radioactivity before performing conventional radiochemical determinations and nuclear studies of the lanthanide fission products. Such a step is required because of the potential hazard involved in handling large amounts of alpha radioactivity directly. I n addition, a t this installation it is necessary to reduce the alpha content of transuranium solutions before transportation to another laboratory for the determination of the lanthanide fission products. Because of its speed and simplicity a recent amine extraction technique ( I , 5 ) may be used for the removal of transuranium elements preferentially. However, some revision of this method for the multiple batch actinide-lanthanide group separation is necessary. Excess hydrochloric acid must be maintained a t all times in the aqueous phase to prevent high losses of lanthanide elements and precipitation of various metal ions. Also, because the lanthanide fission products exhibit slightly different extractability (5), a yield correction based on a single carrier is not representative of the entire group. Use of a mixed rare earth carrier containing lanthanum, cerium, europium, and yttrium has been found quite .atisfactory. The decontaminated
aqueous solution may then be analyzed for total lanthanide fission products by one of several conventional procedures (2,S, 6) with the applicat,ionof a yield correction. EXPERIMENTAL
Reagents. 30% (w./v.) Alamine 336-S-xylene. Dissolve 300 grams of Alamine 336-S in 1 liter of reagent grade xylene. Pretreat the solvent by mixing well for 3 minutes with a n equal volume portion of 1 M hydrochloric acid solution. Centrifuge the organic phase for several minutes and decant it into a glass bottle. Alamine 336-5 is available from General Mills, Inc., Chemical Division, Kankakee, Ill. Mixed lanthanide carrier. Dissolve the appropriate amounts of the oxides or chlorides of lanthanum, cerium, europium, and yttrium to give 5 mg./ml. of each metal in 0.1M hydrochloric acid solution. 1i . 6 M lithium chloride-0.2.M hydrochloric acid solution. Procedure. Dilute the sample solution as much as is practical with 0.1M hydrochloric acid solution. h 1000fold or greater dilution of transuranium solutions is often possible. Pipet a suitable aliquot of the sample dilution into a 50-ml. beaker. Pipet 1 mi. of the mixed lanthanide carrier into the beaker. Evaporate just to dryness on a hot plate. Add 2 ml. of 0.1M hydrochloric acid solution. Evaporate just to dryness. Add 2 ml. of 0.1Jf hydrochloric acid solution and again evaporate just to dryness. Add 10 ml. of 11.6.11 lithium chloride0.2M hydrochloric acid reagent, warm
gently, and transfer to a 50-ml. heavy duty glass centrifuge tube. Add 10 ml. of 3070 Alamine 336-S-xylene and extract for 3 minutes. High speed motor stirrers equipped with glass paddles are excellent for performing the extraction. Centrifuge in a clinical centrifuge for 1 minute. Draw off the organic phase by vacuum and discard. Adjust the aqueous phase to about 0.2.11 i y adding 0.16 ml. of 12M hydrochloric acid solution with a 0.2-ml. graduated pipet. Add 10 ml. of 30% Alamine 336-Sxylene and extract again for 3 minutes. Centrifuge for 1 minute and draw off and discard the organic phase. If more extractions are required, repeat with fresh 10-ml. portions of 30% Alamine 336-S-xylene. Do not add additional hydrochloric acid, however, after the second and third extractions. Wash the sides of the centrifuge tube with about 10 ml. of distilled water and centrifuge for 1 minute. Draw off and discard the organic phase, being careful to lose a minimum of the aqueous phase. Precipitate the mixed lanthanide hydroxides by adding 2 ml. of 6 N ammonium hydroxide. Centrifuge for 3 minutes and decant the supernatant solution. Wash the precipitate by stirring well with about 35 ml. of distilled water. Centrifuge for several minutes and discard the supernatant solution. Dissolve the mixed lanthanide hydroxides with about 10 drops of either concentrated hydrochloric or nitric acid. Analyze the decontaminated solution for total lanthanide fission products by one of the conventional procedures (2, 3 , 6) involving a yield correction. VOL. 37, NO. 3, MARCH 1965
419