VOL. 13, No. 1, 1959

TABLE VII. PRECISION OF M I X I N G POWDERS a

Compos t t ton Measured values o / % e

BaKeL MoKo~ ZnK(x BaL f l , TIK(x

1 1.2 11 1.8 9.4 9.6

2 .8 45 1.8 3.9 6.0

3 2.9 10.2 18.8 5.6 5.6

4 6.6 8.7 4 4 11.0 20.0

A v g 2 9 6.1 6 7 7 5 10.3

Total Average 6 . 7 ~

'~Each value was obtamed f rom three specLmens of the same nommal c o m p o s i t i o n .

Table VIII shows the values of % (robs and the averages. Although the total averages of 2.8% and 0.99% are in reasonable agreement with expected values of ~stat of 3.2 % and 1.0%, the range of % ~ob~ from one laboratory to another is rather larger than would be expected from the chi square tables. One part icipant has suggested that m finding (Xstat from l / N , perhaps N should be taken as the number of pulses entering the detector rather than the number registered by the circmts. This is certainly rea- sonable if the detector is operating m a partmlly saturated con&tion, but detailed measurements are not available for complete interpretation of the data at this time.

Conclusion

The results from this first series of tests in&cate that x-ray data from a number of laboratories may be treated by accepted statistical methods to obtain the standard de- watlons for the various factors affecting precision. The contributions to precision found in these tests should not be regarded as limits on the method but only the expected values for ordinary day-to-day practice. Some of the con- tnbutions to precision will not always occur; for instance it may not be necessary to reset the operating conditions between specimens or to reposxtion the spectrometer be- tween readings.

Estimation-of-composition experiments were not par- t icularly sigmficant in these tests because the specimens

7

TABLE VIII. MEASURED VALUES OF STANDARD DEVIATION (% O'obs) AT TOTAL COUNTS OF 10 3 AND 10 ~

B a K , M o K , Z n K z Lab

. ~ - - ~ 10 3 10 ~ 10 ~ I0 ~ 10 3 I0 ~

1 1.7 5 .3 .1 4.0 0

.5 1.5 1 9 7 1 6 .5

3.6 1.4 .2 7 1.0 .1

2 6 .2 8 4.8 .6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 .8 .7 3 9 .9 5 1 .9

3.8 2 6 5 0 .5 3.1 1.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 4 3 1.2 .4 11.8 1.05

2 5 2.3 1 2 .8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

16 1 6 8.4 3.2

.5 .7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 2 0 1.1 1.0 .91 1.7 1.3

1.5 1.0 6 1.2 1 3 1 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A v g 2 3 1.25 2.3 .9 3.8 .83

Total A v g 1 0 3 ~ 2 . 8 ~ 104 = 0 . 9 9 %

were not realistic from a practical point of view. Almost any judicious choice of standards would lead to far better results than the 2% composition errors observed here.

In checking the detector statistics, the average observed values of standard deviation agreed rather well with the expected values from the (r z ~ N relation. However, the spread m observed values from laboratory to laboratory indicates that there may be aspects of detector statistics that are not ful ly explained as yet.

Future tests wdl be designed to explore some of the contributions to precision in more detail and to set up standard practices that wdl yield data of known depend- ability.

Literature Cited

1. J. Sherman, Proceedings of the 6th Annual Denver Re- search Conference on Industrial Alplicattons o/ X-Ray Analysis, p. 231, Denver, Colo., 1957

Submitted March 3, 1959

v

Spectrophotometric Determination of Nickel With Mercaptoacetic Acid

Philip C. Christopher and Harry W. Hamme

Chemstrand Corporation, Pensacola, Florida

Abstract

A new method for the determmatlon of mckel wi th mercaptoacet*c acid ~s described. The method *s senstttve down to 0.001 mg mckel per ml of solutLon and Beer's law xs followed up to concentrations of 0.7 m g per ml. The color is stable for several hours and the p H of the solution is not critical as long as it is m the alkahne range. Pre- h m m a r y separatmn of mckel by the d*methylglyox*me extract*on method ehmmates interferences. Sodmm citrate, when present in the solut*on before color development, prevents interference f rom metals extracted wath the mckel dur ing the separation and enhances the sensttxwty of the method.

Introduction

Mercaptoacetic acid has been used successfully m ana- lytical chemistry for separation and determinatton of sev- eral ions. I t is especially useful for colorimetric determina- tions because it is colorless and dilute solutions of it can convemently be prepared by diluting it with water. I t has also the advantage in that it acts not only as a color de- veloping agent, but as a complexmg agent as well, prevent- mg formation of the hydroxides of many metals when use

8 APPLIED SPECTROSCOPY

of it is made in alkaline media. Trepka-Bloch (1) using mercaptoacetic acid as a complexmg agent succeeded in separating nickel as the dimethylglyoxlmate from iron, co- bait, copper, silver, lead, arsenic, and bismuth. Nickel i t - self forms two colored complexes with the reagent, one in acidic solution which has a green color, and one in alkaline solution which has a violet color. The complex formed in alkaline solution and having a spectrum with a maximum at 510 m/z is used for this determination.

Due to the reactxwty of mercaptoacetlc acid with other elements and the alkaline conditions employed, most of the common metals interfere with the determinatton. Prehm- mary separation of nickel by extracting i t as the dimethyl- glyoxlmate with chloroform (2,3) or benzene (4) , elimin- ates interferences. The nickel can be transfered to the aque- ous phase by shaking with dilute hydrochloric acid, or the organic solvent can be vaporized and the organic mat ter destroyed by wet oxidation. The second method was found more feasible by the authors. Heat ing the residue with perchloric acid (60-70 percent) for a few minutes, after solvent volatlhzat~on, clears the solution and prepares ~t for color development.

Although prehminary separation of the nickel from other metals is necessary, the method is not time consuming due to the fast extraction method available for separation. The method has the advantage in that the p H is not criti- cal for a wide range and the color is very stable. I t can be especially useful in cases where metals wMch interfere with the dimethylglyoxime method are present.

Experimental

Apparatus The Beckman Model B and Model DK-2 spectropho-

tometers with 1-cm and 5-cm cells were used for absorp- tion measurements and spectra studies.

A Beckman Model G p H meter was used for p H measurements.

Reagents Mercaptoacetic a c i d s o l u t i o n . Prepare by diluting 2

ml mercaptoacetlc acid (Eastman Kodak, or equivalent) to 100 ml with distilled water.

Sodium citrate solution. Dissolve 25 g reagent grade salt per 100 ml of distilled water.

Boric acid in potassium chloride solution. Dissolve 12.4 g reagent grade boric acid and 14.9 g reagent grade potassmm chloride in one liter of distilled water.

S tandard nickel solution. Prepare by dissolving 0.4050 g of uneffloresced crystals of NiCle6"H,_,O in 1000 ml distilled water.

Sodium hydroxide solution. Prepare by dissolving 2.0 g reagent grade sodium hydroxide pellets in 100 ml of distilled water.

Dimethylglyoxlme solution. Prepare by dissolving one gram of dlmethylglyoxime per 100 ml of methyl alcohol.

C h o i c e of pH

The stability of the nickel-mercaptoacetic acid system as a function of pH is presented in Figure 1. In this graph the Absorbance measured at 510 m# is plotted versus the pH. The absorbance is due to more than one nickel com- plex in some portions of the p H range presented. Nickel forms two complexes with mercaptoacetic acid. The one begins to form at about p H 3.0, and reaches maximum

031

w02 z

00

0 o'1

OC 0 o 6'o ~o Ib I~ J~

I /- /

I/

/I I /

/ \

\

20 40 pH

F ic . 1. ABSORBANCE AT 510 MILLIMICRONS D U E T o T H E

N I ( I I ) - M E R C A P T O A C E T I C ACID COMPLEXES VERSUS P H

The solutions were prepared by adding, to a mckel solution con- taming 10 mg mckel, 25 ml o:f the potassium chloride-boric acid solu- tion, 10 ml off the mercaptoacetlc acid solution, and enough sodmm hydroxide solution until pH reached the desired value. The mixture was dduted to 100 ml.

concentration at about p H 5.0. Further increase of p H causes a decrease in concentration of this complex. A sec- ond complex begins to form as the concentration of the first complex begins to decrease, and reaches maximum concentration at about p H 8.0. Further increase in p H has no effect on the stabili ty of this second complex. This is the main reason this complex was used for this determina- tion. The dotted lines in Figure 1 represent the theoretical decrease in concentration of the first complex, and the m- crease in concentration of the second complex as the p H goes from acidic to the alkaline range. I t seems that the two complexes are found in a state of equilibrium which depends on the p H of the solution. This view is supported by the spectra of the complexes presented in Figure 2 at different p H values. From studies of the composition of the two complexes, the green complex formed in acidic solution corresponds to the mono-mercaptoacetate, and the violet complex :formed in alkaline solution to the bis- mercaptoacetate.

06

0S

0 4

o = z ~o3 0~ o ~o2

OI

pt.; 578

,H 472

PH • 62

0 4,0 4 ; 0 S;O 5'50 600 700 WAVE LENGTH-MILL I M I C R O N S

FIG. 2. ABSORPTION CURVES OF T H E N I C K E L ( I I ) - M E R -

C A P T O A C E T I C A C I D C O M P L E X E S AT D I F F E R E N T P H V A L U E S

(Obtained with the Beckman Model DK-2 Spectrophotometer )

The solutions were prepared by adding, to a mckel solution con- taming 25 mg mckel, 25 ml of the potassium chlor~de-borlc acid solution, 10 ml of the mercaptoacet~c acid solution, and enough sodmm hydroxtde soluuon untd pH reached the desired value. The mixture was dduted to 100 ml.

VOL. 13, No. 1, 1959

From Figure 1, although pH is not critical above 8.3, a p H of 11.5 is recommended in the procedure. This was done to eliminate interference from iron which may be extracted with nickel during the separation. Iron forms a violet colored complex with mercaptoacetlc acid in alka- line solution similar to that formed by nickel under the same conditions; however, when the p H is higher than 11.0 the iron complex fades to colorless. When in this colorless form it gives slightly higher results only when it is present at concentrations ten to fifteen times that of nickel. Under usual conditions the amount of iron extracted with nickel is much less than this ratio.

Sample Preparation The separation of mckel by the dimethylglyoxime ex-

traction method has been described m detail elsewhere (2,3,4). The method varies slightly, depending on the nature of the sample, but it is usually fast and quantitative. The procedure used for these studies is essentially a slight modification of the Sandell method (3) . To the acidic sample solution containing Ni (II) in a separatory funnel, 25 ml of the sodium citrate solution and 10 ml of the dl- methylglyoxime solution were added. To the mixture, con- centrated ammonium hydroxide was added dropwise until alkaline, and then a few drops in excess. The mixture was cooled to room temperature and let stand for a few min- utes. The nickel dimethylglyoxlmate was then extracted with chloroform, and the combined chloroform extracts used for the determination as described below.

Procedure Evaporate the organic solvent to dryness being careful

not to lose any of the residue at the end of the evaporiza- tlon. Add 2 ml of perchloric acid (60-70 percent) , and cover the beaker with a watch-glass. Heat on a hotplate until the organic mat ter is decomposed and the solution is clear. Cool and dilute to about 20 ml with distilled water. To this solution add 10 ml of the sodium citrate solution, 25 ml of the potassium chloride-boric acid solution, and 10 ml of the mercaptoaeetlc acid solution. Mix well after each addition. To this mixture add sodium hydroxide solu- tion dropwise until the p H is 11.5. Transfer the solution into a 100-ml volumetric flask and dilute to the mark with distilled water. Mix and measure the absorbance with 1-cm or 5-cm cells at 510 m~. Determine the concentration of nickel from a calibration curve prepared from the standard nickel solution following the same procedure used for the sample.

T A B L E I . R E S U L T S O B T A I N E D F O R N I C K E L W H E N P R E S E N T

A L O N E A N D W I T H O T H E R M E T A L S

Metal Ntcl~el Raho Metal Present, Present, of Metal Nlcl~el Pound

Present Mg Mg to Nickel Mg %

None - - 5.00 - - 4 97 99.5 None - - 5.00 - - 5.06 101 0 None - - 0 50 - - 0.51 102 None - - 0.50 - - 0.52 104 Fe 50 5 00 10'1 5 11 102.2 Fe 75 5.00 15.1 5 19 103.8 C u 100 5.00 20.1 4 9 6 92 2 Cu 100 0 50 200 1 0 51 102 Mn 50 5.00 10.I 5.03 100.3 Mn 50 0.50 100.1 0.47 94 Cr. 50 5.00 10"I 4.95 99 6 Cr 50 0.50 100'1 0.50 100 Mg 100 5.00 20 '1 4.87 97 4 Mg 100 0.50 200 1 0 48 96 Z n 250 5 00 50.1 5 01 100 2 Zn 250 1.00 250 1 0.99 99.0 V 50 5.00 10:1 5 03 100 6 Mo 50 5.00 10 1 4 9 9 99 8 Mo 50 0 50 100 1 0.49 98 Co 50 5.00 10 1 5 17 103.4

TABLE II. C O N C E N T R A T I O N O F M E T A L S P R E S E N T W I T H

N I C K E L I N A L U M I N U M A L L O Y S

Metals W t Grams W t %

A1 0 5000 74 35 Fe 0.0075 1.12 St 0 0050 0 74 Zn 0 0325 4.83 TI 0.0350 5.20 V 0 OO5O 0 75 C r 0.0135 2.01 M n 0 0200 2.97 Mg 0.0140 2 O8 C u 0 0400 5 95

Evaluation of the method

To evaluate the reproducibili ty and accuracy of the method, composite samples made up of different metals and of known nickel content were analyzed. The samples were prepared by mixing standard solutions of different metals with the appropriate amount of standard nickel solution. The nickel content of each sample was adjusted to give a final concentration within which Beer's law is valid. The mixture was heated to dryness and the residue taken up in hydrochloric acid. Separation and determina- tion of nickel followed, as described under sample prepara- tion and under procedure.

Table I represents data obtained for nickel when present alone and with a single metal at concentrations tenfold or higher than that of nickel. Table III represents data ob- tained from composite samples made up of different metals at concentrations found in various aluminum alloys. The concentration of each of the different metals m each sample was kept constant, as given in Table II, and the nickel content varied as shown in Table III.

T A B L E I I I . R E S U L T S O B T A I N E D F O R N I C K E L W H E N P R E S -

E N T I N A L U M I N U M A L L O Y S U P TO 1.5 P E R C E N T

Nwkel % of Nwkel Recovered Added, Mg Sample W1. Mg %

0 50 0 074 0.50 100 0 50 0 074 0.47 94 2.50 0 3 7 0 2 5 1 1 0 0 4 2 50 0 370 2.47 98.8 5 00 0.740 4.99 99.8 5 00 0 740 5 06 101 2

10.0 l 49 9 99 99.9 10 0 1.49 9.97 99 7

Discussion

The use of mercaptoacetic acid method for the nickel determination could be adopted for any sample from which nickel can be separated quanntatlvely. The dimethylgly- oxime method is recommended for separation, since it has been used for practically any type of sample. Small amounts of other metals are always extracted with nickel by this method and copper accompanies nickel to a greater or less extent. Interference from these metals is eliminated by the mercaptoacetlc acid method when sodium citrate is present. Copper was found not to interfere up to concentrations ten times that of nickel, even when separation by extrac- tion does not precede the determination. Sodium citrate has no fleet on the nickel-mercaptoacetlc acid system, ex- cept that it enhances the absorbance by about five percent when present at the concentration stated in the procedure. Data in Tables I and III show that the method can be used on samples composed of many different metals without any serious interference.

Beer's law is followed up to concentrations of 0.7 mg nickel per ml of solution, which gives an absorbance of 1.2 when measured m 1-cm cells.

I 0 APPLIED SPECTROSCOPY

Following Ayres treatment (5), the optimum range of concentration for best accuracy extends from 0.04 to 0.5 mg nickel per ml of solution when 1-cm cells are used and optical density is measured at 510 m~. The relative analysis error per 1% absolute photometric error over this concen- tratlon range is 3.0%.

The color reaches maximum intensity readdy and is stable for at least 24 hours. The stability decreases with the presence of other metals in the solution. The extent of decrease depends on the type and concentration of the metals present, but the stabihty is never less than an hour.

Literature Cited

1. E. Trepka-B'loch, Chem. Analyst 43, 63-65 (1954) 2. O. R. Alexander, E. M. Godar, and N. J. Llnde, Ind.

Eng. Chem., Anal. Ed. 18, 206 (1946) 3. E. B. Sandell, and R. W. Perlich, Ind. Eng. Chem., Anal.

Ed. 11,309 (1939) 4. W. C. Johnson, and M. Simmons, Analyst 71, 554

(1946) 5. G. H. Ayres, Anal. Chem. 21, 652 -949)

Submit ted Augus t 8, 1958

Y

Determination of Zirconium in Plutonium By Ion and Spectrography

Exchange

Roy Ko

General Electric Co., Richland, Washington

Abstract

The spect rographic analysis o:f p l u t o m u m :for ~mpur~ty elements is not w i thou t dxfl~cultles due to the complex emxsslon spec t rum o:f p l u t o m u m and to its toxici ty. Separat ing the p lu ton ium :from the elements o:f interest would reduce the spectral mter:ference and also the heal th hazard In the determinat ion of z~rcomum m p l u t o n m m reported here, the p l u t o m u m was separated :from z~rcomum by re- t aming the p t u t o n m m as a mt r a t e complex on the amon exchange resin, Dowex-1, whde collecting the z~rcomum m the effluent and wash solution. The combined effluent and wash solutions were evap- orated to a small volume and the result ing solution analyzed :for z l r conmm by graphi te spark excxtatlon using cobalt as the in ternal s tandard. Concent ra t ions as low as 20 ppm of z~rcomum xn p l u t o m u m have been determined wxth a precision o:f ± 1 7 % (95~ c.1.) :for a single measurement .

Introduction

In the spectrochemical analysis of plutonium, the tox- icity and complex emission spectrum of the element must be considered. The carrier distillation method of Scribner and Mullin (1) mimmizes the spectral interference, but still presents a considerable health hazard. A preliminary separation of plutonium from the elements of analytical interest would reduce the health hazard and at the same time eliminate the spectral interference.

Solvent extraction has been used for such a prehmin- ary separation. Cupferron (ammonium phenylnitrosohy- droxylamine) extraction (2) and T T A (thenoyl trlfluor- oacetone) and hexone (methyl isobutyl ketone) extraction (3) have been applied at Hanford to the spectrographic analysis of plutonium.

Ion exchange should also be apphcable to such separ- ations. The Canadians at Chalk River have made studies of the behavior of Pu IV on the anion exchange resin, Dowex- 1. ( 4 - 7 ) They have also used anion exchange techniques for the separation of plutonium from fuel element dis- solver solutions (8- 10). Dowex-1 has also been used at Hanford (11) for the separation and purification of plu- tonium. Amon exchange separation as an analytical method for the determination of impurity elements m plutonium should, therefore, prove feasible. A preliminary study with zirconium as the impurity element was made and is re- ported here.

Reagents and Materials

HC1, H N O ~ - - C . P. reagent redistilled from all quartz apparatus

HF-- reagent grade Fluorolube so lu t ion- - l% in petroleum ether Cobalt solution--0.01% in 4 M HNO3 Dowex 1 X 4, 50-100 mesh ¼" x 1" Nattonal Carbon spectroscopic graphite elec-

trodes Eastman spectroscopic plate, type III-0

Apparatus

For the ion exchange separation, 3 ml of resin are packed into 6 mm Ld. glass colums 4" long. The rate of sample and wash solution through-put is 3 m l / m i n / c m 2.

For the spectrographic analysis, a Bausch and Lomb large L t t r o w quartz spectrograph is used in conjunction with a Batrd source power unit. High voltage condensed spark excitation at 5 amperes (primary) with an induc- tance of 40 microhenrys and a capacitance of 0.0025 mic- rofarads is used. Thir ty seconds exposure, extra small col- hmator mask and a 10 micron slit are used. The excitation is made in a glove box set at right angles to the regular Bausch and Lomb arc stand track. The beam is directed into the spectrograph by a lens and mirror system.

Procedure

Fifty mdligrams of plutonium are cautiously dissolved in 1 ml HC1. Then 0.1 ml cobalt solution, 1 drop 0.1 M HF, and 1 ml HNO3 are added, the solution evaporated, taken up in 1 ml 7.2 M HNO3 and passed through Dowex 1 prevxously treated with 6 ml 7.2 M HNO3. The column is washed with 6 ml 7.2 M HNO~, the combined effluent and wash solutions evaporated, taken up in 0.5 ml HC1, 0.1 ml transferred to a pair of graphite electrodes whose top surface has been treated with a drop of fluorolube solution, and the evaporated residue sparked in the glove box en- closed arc stand. The zirconium 3438.2A and the cobalt 3453.5A hnes are densitometered, converted to intensity ratios using the emulsion calibration curve, and the zir- conium concentration determined from the working curve.