Replicability and Breci,ion of the Quantitative Mehod Qf fleterminatiorn Quantity, Nenn IM1. 1 2 3 4 Values
?able 111.
0.133 0.132
0.136 0. I33
10
0 265 0.390
15
0.260 0.2ri7 0,395
0,400
20 20
0.3313
5
5
Q .264
IO
I
15
0.538
0.395
0,533
0,533
0.131 0.133
0.133 0.132
0.266 0.265 0,404
0.268
0.401 0.532 0.533
0.266 0.399 0.400 0.527 0.533
0.133
0.266
0.3‘30
ciiffurcnt fl:ivo~oids cvcn within the smo group may vary corisicterihly. The use of etmidards for cotiipamtive purposcs or tho prepamtion of standard plots for individual compounds is ncccssary. The absorption mnsiniuin for the isonicotinyl hydrazones \vas t)rt\wen 404 : i n d 408 m p . For practicd use. n ninsin i i i ~ nof 40.5 1 1 1 \\’as ~ (.I’ 1OSCIl.
0.532 LITERATURE CITED
L,,Zweig, G., “Manual of Paper Chroniatography and Paper Electrophoresis,” 2nd ed., p. 327, Acadomic Press, New York, 1958. (2) Gage T. B. Douglass, C.D.,Render, S. H I~NAL..&”I. 23,158? (1951 (3)&&man, T. A,, “Moderne Met oden der Pflanzenannlyso” Vol. 3, pp, 4p50-98, Springer Verlag, Berlin, 198.5. 141 . . Pew. J. C.. J. rim. Chew. SOL 70, 3031 (1948). ’ (5) Shih, M. Goodhart, R. S. “The Flavonoids In Biolo and Medjcjne,” Xutrition Monoprng Seriea, NO. 2, Jnnunry 1956. ( 6 ) Smith, L. L., Focll, T., AISAL.CHEM. 31 lOX(1050 ( i )l‘inbcrger, &, J., Ibid., 27,7138 (1955). (8)Wcichselbaum, rF. E., Margraf, H. ?‘., J. Clin. Endocnnol. and ililclaboltsm 17,959 (1957). (9) Wender, S. H., Gage, T. B., Science 109,287 (1949). (10) Willstaetter, R., Ber. 47, 2874 (1914). (11) Wilson, C. W.,J . Am. Chem. SOC. (1) Block, R. J,, Durrum E.
Table IV. Absorptivities (E) of Isonicotinic Acid Hydrazones of Several Flavonoids (.4t 40; mp, incubated a t 50” C. for 150
minutes) Group E 3 260 Flavonol glycoside 8,700 Hesperidin Flavanone Hesperctin Flavnnone 10,900 Narin ‘n Flavanone 10,%IO Flavonol 4,700 2,3-Di&droquercctin Hesperidin methyl- Chalcone i ,GOO chalcone Compound Rutin
I
~-
The smsitivity of detection of certain flavonoids on paper chroiiiatograms with the strong IT\“ reagent extends to levels as low as 5 7. Other reactions involved in the scheme of analysis must necessarily :tiso lie applied on R micro Ecnle. Evidence for tlic precision anti reproducibility of the quriiititative iiiet!iod
is givct; i n Tnble 111. Rscellcnt ngrcenmt on n day-today basis has been observed. The effect of temperature and time on color development for hesperetin is shown in Figure 2. Full color equivalence cannot be reached under the proposed reaction conditions. Varying the temperature markedly affected the rate a t nliich the zbsorbancc increased. However, the absorbance obtaincd under these conditions foIlows Bccr-Lambert’s Iau. in a range from 0 to 25 y of licsperetin per Inl, of rcngcnt. Siiiiilnr standard plots hnve bccn obtained with other compounds. The reaction mas not affected by light. There was no evidence, when the reaction \vas carried out in s n open tube a t room temperature, that the presence or ahscnce of oxygen affected the reaction in any nay. I n this laboratory the reaction takes place in tightly stoppered tubes to prevent evaporation of the solvent. .4bsorptivitics of various eompounds arc prescntcd in Table I V ns evitlence that color equivalences of
i;
F.,
61,3303-6 (1939).
RECEIVED for review August 3, 1959. .4ccepted October 21, 1959. Work supported in part by a grant from tho U. S.
Piitrlic Health Service, C-Z383(C4).
EN D. MOLT Argolane Nafionaf laboratory, Lemont, I f f .
A method i s described for the separation of microgram quantities of silicon from plutonium and other metals for colorimetric determination. The silicon is volatilized as the tefrafluoride from a perchloric acid solution in an allplwtonium, uniformly heated still. The distillate i o absorbed in a solution of boric and molybdic acids, in which the molybdenum blue CO~OPis developed. Most of the silicon distills over during the fou& and flf?h minutes of the 10ting period. Only 0.05 ml. b trydroflwric acid (50%) ia needed 40r the 0- to 50-y range. Nitric acid i o added with the hydroflwric acid to ensure complete deQ
ANALYTlCAl CHEMISTRY
composition of elemental silicon. Results from the analyses of plutonium, uranium alloys, steels, and phosphoric acid by this technique are presented. HE microdetermination of silicon in plutonium has not been made previously a t this laboratory by the senst tive molybdenum blue colorimetric method because of the spectrophotometric interference and the high alpha radioactivity of plutonium. However, a means of making a preliminary quantitative separation of microgram quanit tities of silicon has been sought, would not only make the colorimetric
method applicable to the analysis of plutonium, but could also be used to eliminate many of the interference difficulties coinmoiily encountered in the analysis of other materials. Certain interferences in the development of the molybdenuni blue color have been pointed out by lLluilin and Riley (fl) and Stricklaxid (f2). Kenyon and Bewiek ( 7 ) demonstrated that absorbance varies inversely with srtlt concentration. I n spite of the lower sensitivity imposed by the presence of appreciable concentrations of foreign ions, however, these and other authors (6, 0, 10) have employed the molybdenum blue method without separation.
Volk and Weintrnub (16) isolated silicon from organic materials by ignition of the sample and fusion of the ash in sodium carbonate. For the analysis of ferrous, ferromagnetic, nickel, and copper alloys, Luke (9) separattd the major constituents from silicon by a carbamate-chloroform extraction; for titanium alloys, Codcll, Clemency, and Nonvitz (4) precipitated the titanium and, without filtration, developed thc molybdenum blue in the supernatant. Xone of these techniques were applictible to thc aniilysis of plutonium. Volatilization of si!icon tetrafluoride R S a means of silicon separation is as old :is the classical gravimetric techniquc in which thc ignited oxides containing silicon dioside arc treated with hydrofluoric and sulfuric acid and evaporated to dryness. I n thc mcthod presented here. silicon tetrafluoride is quantitatively espcllcd from the sample solution in a uniformly heated, all-platinum distillation vcsscl, and the distillate is absorbed in a solution of boric and molybdic acids in which is dcvelopcd the molybtlenuni bluc color.
I1
h
Figure 1.
Distillation apparatus
EXPERIMENTAL
Apparatus. Quantitativn recovery of volatilized silicon tetrafluoride is complicated by its strong affinity tor water to form silicic acid (HISiO,) and for hydrofluoric acid t o form fluosilicic acid (K1SiF6). Not only must it be cspellcd from a dchydrateti soIution at a relatively high tempcrature, but it must remain in such a n environment until it can be directed into a recoverable absorbant. Therefore, the construction and use of the distilling vessel must be such that the silicon tetrafluoride, along x i t h escesq hydrofluoric acid, can be ex elled from the sample solution and discgarged into the absorbant without retention by aqueous condensate along the way. The ell-platinum distilling vessel shown in Figure 1 was listed by the nmerican Plntinum Works (now Baker Platinum Division of the Engelhard Industries, Inc.) in their catalo 75 1049 as a "tubulated Gooch crucifle. I t is no,longer a catalog item but is :Lvailablc upon order from the company. The crucible \vas 2.26 cm. in diameter and 10 cm. long Kith a capacity of 25 rnl. The side-arms were l/l-inch tubing. Heat WEB supplied uniformly to the entire distilling vessel, except the grotruding side-arms, by a 300" to 500 F. heat gun, the barrel of which was extended as indicated in Figure 1. A slot, 6/16 X l * / a inches, cut down from the top of the barrel, permitted the discharge tube of the distillin vessel to be lowered to the position sfown. An aluminum cover was attached to fit over the slot and the remaining opening around the protruding platinum tube was calked with a small piece of aluminum foil. When not completely closed, the opening permitted unheated air to be
paratus. ..\dequato vacuuni \vas provided b a water aspirator or by a small iaphragtn pump or rotary v n p u m pump. fhe apparatus was mounted on a single ring stand. The platinum and plastic veasels were aflked to a +inch section of $/winch stainless steel tubing which was movable up and down tho ring stand rod for insertion into and withdraival from the hcat gun, attached to the ring stand. A 230" C. thermometer was mounted on the movable asseni bly. -4tube furnace, l l / , inches in inside diameter by 4 inches long, was set on end nnd niaintnined a t 300" C. This served to heat tile crucible of the )latinum vessel before distillation bot for dissolution of the snmple and for cvnporation to dciisc fumes of perchloric acid. Reagents. ABSORBIKG SoLuwoss. N i x 200 ml. of saturated boric acid solution with 40 mi, of 10% molybdic acid solution and dilute to 1 liter. T o make the 10% molybdic acid solution, dissolve 25 grams of amInonium molybdate tetrahydrnte in 200 nil, of distilled tvatcr, stir in 20 ml. of concentrated sulfuric acid, and clilute to 250 nil.
All-olatinum
i I
4
HYDROFLUORIC-SITRIC ACID 3 1 1 ~ Combine 5 nil. of hydrofluoric :icid (%I%) with 25 nil. of concentrated nitric acid. .bIMONILTM HYDROXIDE. Saturate 1 liter of .distilled w t e r with anhydrous ammonia gas. SULFURIC-TARTARIC ACID SOLUTION, Make up a 5OO-niI. solution containi? 12.7 ml. of concentrated wifuric act m d S grams of tartaric acid. REDUCING SOWTIOX(2). Dissolve 0:1 grani of l-nmino-2-naphthol-4-sulfonic acid and 0.4 grani of sodium hydroxide in 25 ml. of water; add 5.4 grams of sodium bisulfite and dilute to 50 ml. Store the reagent solutions in polyethylene containers. Procedure. Place a w i g h e d sample in the crucible of the distilling vessel and add a minimum amount of dilute nitric or hydrochloric acid, which, when heated with 5.0 ml. of tripledistilled perchloric acid, will decompose the sample. Cranium and plutonium samplcs are conveniently decomposed by successive additions of 1 ml. of 10% hydrochloric acid, 2 ml. of 1 to 1 perchloric acid, and 4 ml. of concentrated perchloric acid. Lower the crucible part of the way into the tube furnace for gentle heating to complete the sample decomposition; then lower it all the way, leaving only the upper rim of the elongated crucible to extend outside the furnace. Heat to co ious fumes of perch!oric acid and withgaw thc crucible from the furnace to cool to room tern erature, (The time distillation schedule of the su!sequent is established on the basis of starting with a dehydrated perchloric acid solution. The perchloric dehydration should not be made in a container other than the di&Alation crucible because of the risk of incomplete transfer of minute quantities of silicon dioxide which TURK
d
V I
'0
10
'
I
I
I
20
30
40
53
'
,
I
60
70
J 80
MICROGRAMS OF SILICON
Figure 2. Distilled vs. undistilled aliquots of standard silicon solutions drawn in a t this point by the updraft within the barrel. This cooled the upper portion of the platinum container, tending to cause premature formation of condensate that could drain back into the joint of the container and escape by capillary action to the outside. With the air-intake damper closed, the gun was operated a t a volttige rcgulated to produce distillation a t the desired rate. The outlet platinum side-arm was connected by a snug slip joint to I/,inch Teflon t,ubing extending into the absorbin solution in a 15-dram transparent pyastic vial fitted with a polyethylene stopper, The inlet end of the Teflon tube was preshapcd to fit the latinum tube by gently heating the atter in a cool Bunsen flame and gradually slipping it about l/, inch into The bends the heated Teflon tubin in the inlet and outlet Te8on tubes were formed by heat treatment. The Teflon plu in the sto cock on the suction a a s t was grooveifto give precise control of the rate of air flow through the a g
P
I
VQL 32, NO. 1, JANUARY 1960
b
125
adhere to the container walls during dehydration.) Place the platinum distilling head on the crucible and wrap the joint with a 3 X 5 inch sheet of aluminum foil, crimping it to a snug fit to minimize air movement through the joint during distillation. Suspend this distilling vessel on a stainless steel wire within the barrel of the heat gun, a s shown in Figure 1. Place the cover over the slot in the barrel and calk the annular opening around the outlet tube with a fragment of aluminum foil. Connect the platinum discharge tube to the Teflon tube of the absorption vessel to which have been added 25 ml. of absorbing solution. Make connection from the absorption vessel to the grooved stopcock. Start the air flow a t about 150 ml. per minute. Through a small Tygon funnel, introduce 0.3 nil. of the hydrofluoric-nitric acid mixture into the distilling vessel. (The purpose of the nitric acid is to effect complete decomposition of elemental silicon n hich may have remained unattacked by previous treatment in the absence of the hydrofluoric-nitric combination.) Start the distillation by energizing the heat gun through a n alarm timer set for a 10-minute OKinterval. During the distillation, note the thermometer readings and look for occasional droplets of distillate being swept through the translucent Teflon tube into the absorbing solution. Beads of nitric acid distillate, 1%-hichare visible a t first, gradually disappear; then during the last 5 minutes of heating 3 or more drops of perchloric acid, n hich have accumulated in the platinum outlet tube, migrate quickly through to the absorbant solution. The temperature of the heated air in the gun barrel near the top of the still should increase to about 200" C. When the timer alarms and shuts off the heat gun, disconnect the absorbing vessel, first from the platinum and then from the suction tube. Remove the plastic stopper and rinse the Teflon tube 11ith miter. Sdjust the p H of the absorbant solution to 1.2 to 1.3 with silica-free animonium hydroxide. Allow 10 minutes for the formation of the beta form of silicomolybdic acid. Yolk and JTeintraub (15) reported that the p-silicomolybdic acid forms completely within 4 minutes and remains stable for a t least 6 minutes, after which it slowly changes to the alpha form. Jean (6) confirmed the transformation from the beta to the alpha form by polarographic measurements and demonstrated the acceleration of this process by increasing the time, temperature, or pH. Add 5 ml. of the sulfuric-tartaric acid mixture and stir. Without the presence of tartaric acid, the niolybdenum blue later developed by reduction of the silicomolybdic acid is not stable but continues to grow denser with time, because of the slow reaction of excess molybdic acid with reducing solution. The sulfuric acid is added to produce high acidity a t the time of reduction, which, according to Nullin and Riley (11), is necessary to avoid interference 126
ANALYTICAL CHEMISTRY
DISCUSSION A N D
0.I o'2 o
i '
0
I , 0 2'
l 0.4
,
I
0.6
.
I
0 8
ML. OF HYDROFLUORIC ACID (50%)
Figure 3. Variation of absorbance with amount of hydrofluoric acid used in distilling 10 y of silicon
by the excess molybdate reagent. Other authors have found that interference by phosphorus and arsenic is also prevented by high acidity a t the time of reduction (1, 6). Immediately after acidifying. add 0.8 ml. of the reducing solution, dilute to 50.0 ml., and alloiv to stand for 20 minutes. A comparison of five usable reducing reagents by Mullin and Riley revealed t h a t l-amino-2-naphthol-4-sulfonic acid was the most rapid one, although not the most sensitive. For their work on sea n-ater they used p-methylaminophenol sulfate (nietol) which was more sensitive but required 2 hours for complete color clevelopnient ,
Table I. Effect of Excess Perchloric Acid in Absorbing Solution on Absorbance Per Cent 1 to 1 HC104, Absorbance, hlaximum of M1. 1 Cm. 0.0 0.782 100 0.5 0.755 97 1 .o 0.725 93 1.5 0,684 88
Transfer the colored solution to a l-cm. cell if in the 0- to 50-7 range, or to a 5-cni. cell if in the 0.0- to 10.0-7 range, and measure the absorbance a t 815 mu using distilled n-ater as reference. Subtract the absorbance of a corresponding blank carried through the entire procedure and calculate the silicon content as follows: P.p.m. of Si =
net absorbance X factor grams of sample
where the factor is the average ratio of micrograms of silicon per net absorbance obtained on aliquots of a standard silicon solution.
RESULTS
Sulfuric acid, used for volatilizing silicon tetrafluoride in gravimetric procedures, was found to be unsuitable for this method. Precipitated sulfates within the crucible bumped and spattered a t elevated temperatures. Furthermore, quantitative volatilization did not occur unless the volume of the hydrofluoric acid was considerably larger than that of the sulfuric. Such amounts of hydrofluoric distillate in the absorbing solution prevented color development. The quantity of perchloric acid which distills into the absorbing solution is a n important factor in the final measurement of the distilled silicon. Either too much or too little causes low results. If little or none comes over, some silicic and fluosilicic acids niay remain in condensate deposits formed earlier in the upper portion of the still and in the platinum outlet tube. On the other hand, excessive amounts of acid in the absorbing solution tend to repress the formation of the p-silicomolybdate coniplex, even though the p H is adjusted thereafter. The effect of the initial acidity of the solution on the final absorbance is demonstrated by the data of Table I, taken from four aliquots of standard solution, each containing 50 y of silicon. Each was pipetted, without distillation, into 25 ml. of absorbing solution; then 0.3 ml. of hydrofluoric-nitric acid mixture and varying amounts of 1 to 1 perchloric acid i-iere added, the pH was adjusted, and the colors were developed according to the procedure. I n vieir of these considerations, the ultimate criterion for the proper conditions of distillation (temperature, length of heating time, and rate of air flow) is the amount of perchloric acid brought over, a n estimate of nhich can be made from the amount of ammonium hydroxide required for p H adjustment. Distilled aliquots of standard solutions produce a ratio of micrograms of silicon per absorbance about 37, lower than corresponding undistilled ones. Figure 2 illustrates this difference for aliquots in the 0- to 50-7 range. This small discrepancy is attributed not to incomplete transfer of the silicon but to the small excess of acid in the absorbing solutions of the distilled aliquots. The average amount of distilled perchloric acid necessary for the complete transfer of silicon n-as about 0.25 nil., which, according to Table I. reduced the absorbance about 3%. This involves no error in the determination of unknon n quantities of silicon, provided the micrograms of silicon per absorbance factor used in the calculation are derived from distilled standard samples. The optimum amount of reagent grade hydrofluoric acid (50%) for each
distillation was 0.05 ml. This is a 1TOfold excess over the stoichiometric requirement for 50 y of silicon. Figure 3, showing the variation of absorbance with volume of hydrofluoric acid, indicated that any volume from about 0.05 t o 0.6 ml. was adequate, the lower quantity contributing slightly less to the reagent blank. Above 0.6 ml. the density dropped off sharply because of the bleaching effect of the fluoride in excess of t h a t complexed with boric acid. Tests were made to determine minimum boric acid requirements for complexing the fluoride in distillates of 10-7 aliquots. Curves A and B in Figure 4 show the variation of absorbance with volume of saturated boric acid solution for 0.05 nil. and 0.10 ml. of hydrofluoric acid, respectively. According to curve A , 5 nil. of boric acid solution are adequate for 0.05 ml. of hydrofluoric acid. The slope of the merged A and B curves represents the reagent blank of the boric acid. This same slope is noted in curve C representing results obtained in the absence of a n y fluoride. The lowered
absorbance values of curve C demonstrate the contribution of hydrofluoric acid to the enhancement of color b y assuring complete solution of all silica (2,3). hlost of the silicon was transferred from the sample solution to the absorbing solution during the fourth and fifth minutes of heating. Figure 5 illustrates this pattern of distillation during the 10-minute heating period for both a 5and a 50-7 sample. I n each of these experiments five separate 25-ml. ab-
sorbing solutions were prepared before beginning the distillation. rlt the end of each 2 minutes of heating a vial of fresh solution was connected to the Teflon delivery tube. Pertinent to these patterns were the observations that the temperature within the heat gun barrel reached 200" C. within about 3 minutes and droplets of perchloric acid distillate appeared in the discharge tube after about 5 minutes. Table I1 shows the results obtained on a group of 0.3-gram plutonium samples
Table II. Micrograms of Silicon in 0.3-Gram Samples of Plutonium
ildded
in Spike 5.0 5.0 10.1 10.1 10.1 30.2 30.2
Ill Pu 0.0 0.0 0.0 0.0 1.6 1.7 1.6
Total
Measured
Deviation
5.0 5.0 10.1 10.1 11.7 31.9 31.8
3.8 5.3 8.5 10.9 11.7 32.4 33.0
-1.2 +0.3 -1.6 +0.8 0.0 10.5 +l.2 *0.8 y
Av. Table 111.
Silicon in Uranium Metal and Alloys
Comparative Results Material Yo.
0.82
; C
G 0.80
D
w 0.78
s m
sic
0.76
K
0.74
2
l
0.05 m l HF e 010 rnl H F A
0
NO HF
-
0.72 C.70
0
2
'
4
8
I
I
10
12
14
C
ML. OF S A T U R A T E D BORIC A C I D
Figure 4. Variation of absorbance with volume of saturated boric acid solution
80 90
t ~
1 7 0 1
5 0 1 of S i
507
Of
s
n
0.07% 0,1270 0.24%
Argonne Sational Laboratory, spectrographic estimate.
c
Los Alamos Scientific Laboratory, spectrographic ( I S ) .
d
Battelle Memorial Institute, gravimetric 14). Silicon in National Bureau of Standards Steels YBS Range, S B S Average, This Method,
Type of Steel
SO.
%
7c
132
No-IT-Cr-V
0.225 to 0.26
0.239
126-4
High nickel
0.192 to 0.198
0.194
130
Lead bearing
0.231 to 0.241
0,237
1.53
co-JlO-IT
0.17G t o 0.200
0,187
0.237 0.239 0.230 0.231 0.224 0.240 0,246 0.245 0.203 0.205 0,239 0.241 0,195 0.197
73
-
Table V.
Silicon in Reagent Grade Phosphoric Acid, 85y0
Si Added, .le HBPOI
Si,
y
As std. soh.
MINUTES
OF
HEATING
-
Figure 5. Distillation patterns
1 0 1.0
1.0 1.u
Measured
Y
Deviation
6.9 6.9 7.2
1.0 1.0 1.0 0 2 4 6 8 1 0
3 . 2 p.p.m.
* Argonne National Laboratory, gravimetric (14).
HgPO4, MI. 1.0
0 2 4 6 8 1 0
d
c
23 23 '74
Table IV.
'
6
b
a
