624
ANALYTICAL CHEMISTRY
Table VI.
Analysis of Methane-Ethane Mixtures with Fluorine
Order of Titration XIixture into fluorine Fluorine into mixture Fluorine into mixture, 1 : 1 nitrogen dilution
Methane Found, % 45.4,46 6,44.2 45.7,49 4,49.4,51.1 49.4
Methane Known, % 48.7 48.7 48.7
Table VII. Determination of Olefins in RIixtures with Chlorine Mixture Propene-isobutane Ethylene-ethane Ethylene-nit rogen
Known Olefin Content, % 59.5 62 4 50.3 75.6 50.5
+
There is an apparent bias of 3% for the ethylene titration as compared t o the iodide method. One source of this is the presence of sodium fluoride in the reaction vessel. This would lead to a positive error because of the presence of hydrogen fluoride in these gas mixtures.
Table VIII. Determination of Fluorine by Titration with Ethylene and by Iodide-Acidimetric Method
Olefin Content Found, % 60.1 63.0 50 0 75.3,76.0 50.2
Iodide Method, Fluorine, Fc
Ethylene Titration, rliiorine. C;
Dynamic 4ystein 43.w 42.6 28.3
49.0,49 5 45.7,44.3 31.0
Static syateni
7.5 1
71 6 69 1
70 4
DETERMINATION O F If ETHANE-ETHANE LfIXTURES WITH FLUOBoth methane and ethane reacted with fluorine, but with different reactant and pressure ratios. A known mixture of methane and ethane was prepared, and the reactant and pressure ratios of the individual gases were used to calculate the expected ratios for the mixture. The results obtained by this method are given in Table VI. This type of determination was a relative one. It was necessary to assume that all the sample gas was either methane or ethane. The presence of a third gas would make the determination impossible. DETERMIKATIOK OF OLEFINSIN HYDROCARBOX MIXTURES. T h e determination of olefins in olefin-hydrocarbon or olefin-nitrogen mixtures was accomplished by titration with chlorine. Several examples are tabulated in Table VII. DETERVINATION OF FLUORINE WITH ETHYLEKE. The determination of fluorine by titration with ethylene was compared with the iodide-acidimetric method for fluorine which is described by Rodden ( 5 ) . Comparative analyses were made both on a dynamic system in which the fluorine concentration could be changed, and on a static system. These results are tabulated in Table VIII. The other major component of these mixtures was nitrogen.
67 0 70 0 72 8
RISE.
a
Av. 7 0 . 8 3.7 370 hydrogen fluoride present.
72.1 74.0 5.7 Not determined on other samples.
+
LITERATURE CITED
Cromer, S., ”Electronic Pressure Transniitter and Self-Balancing Relay,” MDDC-803, Columbia Universit.y, Sew York, June 20, 1944. McMillan, W. .4.,IXD. ENG. Cmix., ANAL. ED., 9, 511-14 (1937). Miller, W.T., dr., J . Am. Chem. SOC.,62, 341 (1940). Pogorshelski, Z . , J . Russ. Phys. Cheni. SOC.,36, 1129 (1904); Chem. Zentr., 1, 667 (1905). Rodden, C. J., Ed., “Analytical Chemistry of the Manhattan Project,” KXES YIII-1, p. 238, Kew Tork, McGraw-Hill, 1950. Taylor Instrument Co., Rochester, N. T.,“Operating Instructions and Parts List for Taylor Pneumatic Transmitter,” 1945. Voldere, G. de, and Smet, G. de, 2. annl. Chern., 49, 661-8 (1910). RECEIVED for review J u l y 67, 1961. Accepted January 12, 1963. Based on work performed for t h e htoniic Energy Commission b y Carbide and Carbon Chemicals Co., a division of Union Carbide and Carbon Carp.
Silicate Analysis by a Rapid Semimicrochemical System RICHARD B. COREY AND kl. L. JACKSON University of Wisconsin, Madison, Wis.
S
EMIMICROMETHODS have been devised for the determination of many elements in a great variety of materials, but little has been done toward devising a complete system of silicate analysis utilizing these rapid techniques. Hedin (3)has presented a colorimetric system of silicate analysis, but the sample size required is too large for several types of investigations. Flow sheets for a new semimicrochemical system of silicate analysis for eight elements (silicon, aluminum, iron, titanium, calcium, magnesium, potassium, and sodium) are given in Figures 1and 2. Two samples are necessary for the determination of the total of eight elements, as it would be extremely difficult t o determine silicon and the alkali metals in the same sample. The products obtained from the various operations are enclosed by a rectangular box. Reagents t o be added and operations to be performed are placed beside the lines joining the boxes. Separations, which in all cases are accomplished by use of a centrifuge equipped with special pointed centrifuge tubes (Figure 3),
are indicated by a branching of the line joining the boxes. The precipitates, which are indicated by double vertical lines, w e written on the left, and solutions on the right. Evolution of a volatile component is expressed by placement of the box which designates that component above the horizontal line connecting it to the rest of the flow sheet. The sj-mbol for the element determined is circled at the place on the flow sheet where it is actually determined. The wave lengths given refer to the filter to be used for the colorimetric determination of that element with the Evelyn colorimeter. Because two samples are used for the analysis, some of the elements may be determined in either sample. The procedures which are generally used for the determination of these elements are indicated by solid lines. Alternative procedures, which are usually applied when only one sample is used (for determination of fewer than the total of eight elements) are indicated by broken lines on the flow sheet. Of the two 0.1-gram samples of each
V O L U M E 25, NO. 4, A P R I L 1 9 5 3
625
It is often necessart- in the quantitative analysis of silicate materials to work with samples which are so small that sufficiently accurate results cannot be obtained by usual gravimetric methods. The time consumed often makes these methods unduly burdensome. . i n analytical system is presented which utilizes rapid semimicrochemical methods in the analysis of silicate materials. The elements determined are silicon, aluminum, iron, titanium, cal-
material used, one ip dcconiposed by treatment with a mixture of hydrofluoric, perchloric, and sulfuric acids, and the other by fusion with sodium carbonatr. The sample decomposed by treatment R ith hydrofluoric, perchloric, and sulfuric acids (Figure 1) is used for the determination of calcium, iion, titanium, sodium, and potassium, Perchloric acid is used with sulfuric acid to decompose any organic constituents which may be present in the sample. ifter evaporation of the mixture almost to dryness, the sample is diluted to 100 ml. -4n aliquot of this solution is used for the determination of calcium. Calcium is separated from iron, titanium, and aluminum by the ammonium hydroxide separation, after which the ammonium salts are destroyed by evaporation with nitric acid. The residue containing the calcium is dissolved in hydrochloric acid and made alkaline with potassium hydroxide. The calcium is then titrated with disodium Versenate by themethodof Schwareenbach and Biedermann (4). Iron and titanium are determined in another aliquot of the same solution by the methods of Yoe and Jones ( 1 0 ) and Yoe and -4rmstrong (Q), while sodium and potassiumaredetermined in still another aliquot with the flame photometer. Magnesium and aluminum could also be run on this sample if dcxiired.
cium, magnesium, sodium, and potassium. Two 0.1-gram samples of each material are used in the analysis, one being decomposed by fusion with sodium carbonate, and the other by volatilization of the silicon with hydrofluoric acid. Analyses of a variety of National Bureau of Standards samples by the new method were in good agreement with the standard values. Eight elements can be determined in eight materials in one day.
The sodium carbonate fused sample (Figure 2) is used for the determination of silicon, aluminum, and calcium plus magnesium, The sodium carbonate melt is dissolved with perchloric acid, and the silica is dehydrated by evaporation of the perchloric acid as suggested by Willard and Cake (8). The silica is separated by centrifugation and dissolved with sodium hydroxide. Silicon is determinedin an aliquot of the resulting solution by the method of Dienert and Waldenbulcke ( 1 ) as modified by Hedin ( 3 ) . The classical ammonium hydroxide separation i p next used to separate
0.100 gm Sample NazCOa HCiO4
l o HCIO4 fumOi Dilute to 60 mi
~
Solution D
Dilute l o 50 ml
pH i o solution
1 NaOH
4
I Eriochrame black T
I HF HeSO4 HCIO. Evaporate almost to drynew
1 Sulfates of K. No, Fe, Ti, Ca, 8 Mgl pH 4.2 b u f f r
1
I
Rkad a 5 2 0
Rad 0 4 0 0
Figure 2.
_______--
J
dilute ~- 5o_ _ml- I .-,---_I
to
J
Read AI 5eo my
PH 10 601U'ion
___-----
L7
~
i Eriochrome black T J
@ with
Titrate verrmate
Figure 1.
Flow Sheet for Sample Decomposed in Hydro5uoric .4cid
y
mp
Flow Sheet for Sample Decomposed in Sodium Carbonate
the calcium and magnesium from the hydroxy oxides of iron, aluminum, and titanium. The precipitated hydroxy oxides are separated by centrifugation, and an aliquot of the clear supernatant liquid is removed. The precipitates are then redissolved. Treatment with a concentrated sodium hydroxide sodium is used to separate the aluminum as the aluminate ion from the precipitated hydroxy oxides of iron and titanium. Aluminum is then determined from an aliquot of this solution by the aluminon method of Smith et al. (6). Iron and titanium could be determined by solution of the precipitate and use of the Tiron method of Yoe and Jones (IO) and Yoe and Brmstrong (9). However, these elements are determined more readily in the other sample.
626
ANALYTICAL CHEMISTRY
Total calcium plus magnesium is determined in the supernatant solution of the ammonium hydroxide separation by the addition of ammonium hydroxide and titration with disodium Versenate by the method of Schwarzenbach and others ( 2 , 5 ) . The calcium equivalent determined in the other sample is subtracted from this valuo to obtain the amount of magnesium in the sample. APPARATUS
For the analysis of eight materials for eight elements, the following are required: eight 30-ml. platinum crucibles; eight 500ml. nickel or platinum beakers; eleven 50-ml. borosilicate glass beakers; nine 125-ml. borosilicate glass conical flasks; eighteen 50-ml. volumetric flasks, one set of 9 to be used exclusively for the determination of iron, titanium, and aluminum, and the other set for silicon; sixteen 100-ml. volumetric flasks, one set of 8 to be used exclusively for Solution A and the other for Solution F; eight 500-ml. volumetric flasks; one (or two for increased speed) 50-ml. Lowy pipet (pipet with 3-way stopcock); pipets delivering 1, 2, 3, 4, 5, and 10 ml. of solution; a suction apparatus for drawing off the supernatant solutions from centrifuge tubes; Evelyn colorimeter with 400-, 520-, and 565-mp light filters; 18 matched colorimeter tubes, one set of 9 tubes to be used exclusively for the determination of iron, titanium, and aluminum, and the other set for silicon; International size 2 centrifuge; 16 graduated 60ml. pointed centrifuge tubes as shown in Figure 3 (can be obtained from International Equipment Co., Boston, hlass.); an air-jet stirrer consisting of a 20-cm. length of glass tubing 3 mm. in outside diameter pulled out to a fine point on one end (filtered air is forced through this tube to mix solutions in the pointed tubes and also to dislodge precipitates from the bottom of the tubes); flame photometer (Perkin-Elmer Model 18 was employed); a 2-liter borosilicate glass beaker to be used as a hot water bath for sets of eight pointed centrifuge tubes. DECOB.IPOSITION WITH HYDROFLUORIC ACID
Reagents. Hydrofluoric acid, 487,; perchloric acid, 60%; sulfuric arid, 1 to 1; hydrochloric acid, 1 to 1. 38 rnm 2 mm
\\-/[
A
2 5 mm
I
60
,k
Figure 3. Special 60-MI. Pointed Centrifuge Tube and Special Rubber Pad for Support during Centrifugation Procedure. A 0.1000-gram finely ground sample Tvhich has been dried a t 105' C. for 2 hours is weighed in a 15- or 30-ml. platinum crucible. Then 0.5 ml. of perchloric acid, 1 ml. of 1 to I sulfuric acid (the sulfuric acid is omitted if the sample contains over 1% of titanium), and, 5 ml. of hydrofluoric acid are added. The partially covered crucible is heated on a sand bath a t about 200' C. until the acids are evaporated almost to dryness. The solution must not boil, or spattering may occur. The crucible is removed from the sand bath and cooled, 5 ml. of 1 to 1 hydrochloric a& are added, and the suspension is diluted to two thirds the volume of the crucible with water. The crucible is
then covered and heated in an air bath (crucible within a larger crucible) PO that the solution boils gently. After 5 minutes of this gentle boiling the residue should be completely dissolved. The solution in the crucible is then cooled, transferred to a 100ml. volumetric flask, and diluted to a volume of 100 ml. This is Solution A. used for the determination of calcium, iron, titanium, sodium, and potassium. The sodium and potassium determinations are run as soon as possible to prevent contamination from the glassxare. If necessary, this solution may alm be used for the determinations of aluminum and magnwiu-n. SODIUM AND POTASSIUM DETER.MINATIO3 S
Reagents. Standard sodium and potassium solutions, stock solutions containing 1 mg. of sodium and potassium per ml. Appropriate dilutions of the solution (dependent on the concentration range of the flame photometer used) are used for calibration of the flame photometer, care being taken that the final solution is 1 to 40 with respect to hydrochloric acid. Procedure. The sodium and potassium contents of Solution A can be determined with a suitable flame photometer. The following is the procedure employed for determination with the Perkin-Elmer PIIodel 18 flame photometer. Approximately 25 ml. of Solution A are transferred to a 50ml. beaker. The flame photometer is standardized for the range of either 0 to 10 or 0 to 2 mg. of sodium per 100 ml., depending on the concentration of sodium in the test solution. The galvanometer reading is obtained for the unknown sample solution, and the concentration of sodium is obtained from a standard curve. The flame photometer is then standardized for potassium, and the same procedure is repeated for the potassium determination. The range of 0 to 2 mg. per 100 ml. is used for samples which contain less than 27, sodium or potassium, and the range of 0 to 10 mg. per 100 ml. is used for samples which contain from 2 to 10% sodium or potassium. The percentage of either sodium or potassium in the sample is (milligrams of sodium or potassium per 100 ml.) X (O.lOO/weight of sample). CALCIUM DETERMINATION
Reagents. Animoniuin chloride; ammonium hydroxide, 1 to 4; nitric acid; potassium hydroxide solution, 10 grams of potassium hydroxide dissolved in 100 ml. of water; standard calcium solution, exactly 0.0100 N . To prepare the standard Versenate solution, a 2.0-gram portion of disodium Versenate (disodium dihydrogen ethylenediaminetetraacetic acid from Bersworth Chemical Co., Framingham, hlass., or Eastman Kodak) is dissolved in 900 ml. of water. The normality of this solution is obtained by titrating a 25-1111. portion of the standard calcium solution, and the solution is then diluted so that the normality is exactly 0.0100, A check standardization is then made. Nlurexide indicator powder, a 0.2-gram portion of murexide (from Eastman Kodak), is mixed with 40 grams of powdered potassium sulfate. Procedure. A 50-ml. aliquot of Solution B is placed in a 60ml. pointed centrifuge tube. Approximately 1 gram of ammonium chloride is dissolved in this solution by agitation with the air-jet stirrer, 3 drops of bromocresol purple indicator are added, and 1 to 4 ammonium hydroxide is dispensed from a buret until the purple end point is reached. The tube is then placed in a hot water bath for 5 minutes to flocculate the R203 precipitate, and cooled, and the volume is adjusted to exactly 60 ml. If the indicator has changed back to yellow, 1 to 4 ammonium hydroxide is added until the purple color reappears. The suspension is mixed thoroughly with the air-jet stirrer, then centrifuged 5 minutes a t 1800 r.p.m. .4 50-ml. aliquot of the supernatant liquid (or less if 50 ml. contains more than 5 mg. of calcium) is transferred to a 125-ml. conical flask, 10 ml. of nitric acid are added, and the solution is evaporated to dryness by boiling on a sand bath. The cooled residue is dissolved with 1 ml. of 1 to 1 hydrochloric acid, and the solution is diluted to approximately 50 ml. This is Solution B for the determination of calcium -45-ml. portion of 10% potassium hydroxide is then added, and 0.3 gram of murexide indicator powder is dissolved in the solution. The amount of calcium in the solution is then determined by titration with 0.0100 N LTersenate solution t o a purple end point. The color change is gradual, so it is best to compare the color a t the end point with a blank. The percentage of calcium in the sample is equal to (ml. of Versenate) x (0.048l/weight of sample).
627
V O L U M E 25, NO. 4, A P R I L 1 9 5 3 IRON AND TITANIUM DETERMINATIONS
Reagents. Hydrochloric acid; sodium dithionite ( NazSz04, from Amend Drug and Chemical Co., New York). T o prepare the Tiron reagent solution, a 4-gram portion of Tiron (LaMotte Chemical Products Co., Department H , Towson 4, h'Id.) is dissolved in water, and the solution is diluted to 100 ml. For the buffer solution, pH 4.7, equal volumes of 1 M acetic acid and 1 M sodium acetate are mixed, and the p H is checked on the glass electrode and adjusted as needed with sodium hydroxide or acid. Standard iron solution, stock solution containing 50 micrograms of iron per ml. Aliquots (2, 4, 6, and 8 ml.) of this solution are taken for the standard curve, and the color is developed as described below, the standard solutions being substituted for the Solution -4aliquot. Standard titanium solution, stock solution containing 10 micrograms of titanium per ml. Aliquots (2 4, 6, and 8 ml.) are taken for the standard curve, and the color 1s developed as described below, the standard solutions being substituted for the Solution A aliquot. Procedure. A 10-ml. portion of pH 4.7 buffer solution is placed in a 50-ml. volumetric flask. The volume is adjusted to approximately 30 ml. with water, and 5 ml. of 4% Tiron solution are added. A drop of 37? hydrogen peroxide is added to the remaining portion of Solution A, and an aliquot containing from 50 to 250 micrograms of iron and 10 to 100 micrograms of titanium is transferred from Solution A to the flask. ( A 5-ml. aliquot is convenient for samples containing up to 670 iron and 1.2% titanium.) The solution is diluted to the mark with distilled water, and approximately 20 ml. of this solution are transferred to a colorimeter tube. The per cent light transmittance of the blueviolet color of the iron complex is read with a 565-mp light filter. The color is stable for a long time. Approximately 3 mg. of sodium dithionite is then added to this same solution in the colorimeter tube to destroy the color due to the iron complex by reduction of the iron. The colorimeter tube is stoppered with a Eo. 3 rubber stopper, and the contents of the tube are mixed gently; the tube iq inverted 4 or 5 times. The per cent light transmittance of the yellow titanium complex is obtained within 10 minutes after the addition of the sodium dithionite with the 400-mp light filter. The concentrations of iron and titanium are then obtained from the standard curves. The percentage of iron in the sample is (micrograms of iron per 50 ml./ml. in aliquot) x (O.Ol/meight of sample). The percentage of titanium in the s:rinpIe is (micrograms of titanium per 50 ml./ml. in aliquot) X (O.Ol/weight sample). SODIUlI CARBON4TE FUSIOY
Reagents. Sodium carbonate, anhydrous. Procedure. A 0.1000-gram, finely ground sample which has been dried a t 105" C. for 2 hours is weighed in a 30- to 45-m1. platinum crucible and approximately 0.75 gram of sodium carbonate is added. The sample and carbonate are thoroughly mixed by rotating the crucible with the fingers. Approximately 0.25 gram of sodium carbonate is then added on top of the mixture. The partially covered crucible is placed in a slanting position on a silica-covered triangle. The low flame of a lleker burner is placed PO as to heat one side of the crucible, and the heat is gradually increased until the melt is liquefied. If the material is finely ground the fusion is usually complete in 1 to 2 minutes, a t which time the flux appears quiet with small curds of precipitate floatclear liquid, and the evolution of gas has e is then grasped with nickel tongs and swirled so as to spread the flus in a thin layer over the sides of the lower half of the crucible, so that any silica adhering to the sides is fused, Heating i y continued under full 116ker flame for 2 or 3 minutes, after which the burner is removed and the crucible swirled again to spread the melt in a thin la)-er to facilitate its solution. SILICON DETERMINATION
Reagents. Sodium hydroxide, pellets; perchloric acid, 60%; hydrochloric acid, 1 to 1 and 1 to 9. To prepare the ammonium molybdate solution, a 150-gram portion of ammonium molybdate tetrahydrate is dissolved in distilled water, diluted to 1 liter, and filtered if cloudy. Standard silicon solution, containing 50 micrograms of silicon per ml., is prepared by fusion of pure quartz crystals with sodium carbonate and subsequent solution of the melt in water. Aliquots (1, 2, 3, 4, 5 , 6, 7 , 8, 9, and 10 ml.) of this solution are taken for the standard curve, and the color is developed as described below, the aliquots of standard solution being substituted for the Solution D aliquots. Procedure. When the crucible from the sodium carbonate fusion has cooled, the cover is placed on the crucible, and 8 ml. of concentrated perchloric acid are added dropwise under the
slightly raised lid. When effervescence has ceased, the lid and sides of the crucible are washed down with a minimum of water, the partially covered crucible is placed in a sand bath on an electric hot plate, and the suspension is evaporated to fumes of perchloric acid. Vigorous boilingmust be prevented or loss of sample may result. When dense fumes of perchloric acid appear, the crucible is covered, and the suspension is boiled gently for 10 minutes a t a temperature a little above 200' C. When the crucible has cooled, approximately 5 ml. of distilled water are added, and the suspension is carefully mixed and heated almost to boiling to dissolve the salts which have solidified on cooling. The suspension is then transferred to a 60-ml. pointed centrifuge tube, and the crucible is rinsed with a wash bottle, the washings being added to the tube. All of the silica does not have to be removed from the crucible a t this time. Approuimately 2 ml. of 1 to 1 hydrochloric acid are added, and the suspension is diluted to exactly 60 ml. with water, thoroughly mixed with the air-jet stirrer, and centrifuged a t 1800 r.p.m. for 5 minutes. The suspension adhering to the air-jet stirrer is washed back into the crucible. A 50-ml. aliquot is pipetted with a Lowy pipet from the supernatant liquid in the centrifuge tube and transferred to another 60-ml. pointed centrifuge tube. This is Solution C which is used for the determination of aluminum and calcium plus magnesium (and also iron and titanium if so desired). The pipetting operation is carried out by means of a Lowy pipet, although an ordinary pipet may be used if care is exercised. The silica in the tube is washed by adding about 50 ml. of 1 to 9 hydrochloric acid, stirring, and centrifuging again a t 1800 r.p.m. for 5 minutes. The supernatant liquid is then decanted by suction and subsequently discarded, since an aliquot has already been taken of the solutes. The silica is then washed from the tube into a 250- to 500-ml. nickel or platinum beaker with a stream from the wash bottle, and the silica adhering to the sides of the crucible in which the dehydration was carried out is also transferred. The silica adheres rather tightly to the crucible, so that it must be loosened with a policeman and washed out with a stream from the wash bottle. The final washing of both the crucible and centrifuge tube is made with warm 5% sodium hydroxide to make sure all of the silica is removed. Approximately 2.5 grams of sodium hvdroxide pellets are added to the suspension in the nickel or p k t inum beaker, and the volume is adjusted to approximately 100 ml. The solution is then boiled for 5 minutes to dissolve the silica. K h e n cool, this solution is transferred to a 500-ml. volumetric flask and diluted to the mark with distilled water. T h b is Solution D, used for the determination of silicon. A 10-ml. portion of ammonium molybdate solution is placed in a 50-ml. volumetric flask, and the volume is adjusted to about 30 ml. with distilled water. Then 5 nil. of 1 to 1 hydrochloric acid are added, and the flask is swirled to dissolve the white precipitate that forms. Finally, a Solution D aliquot containing 50 to 500 micrograms of silicon is added, and the solution is diluted to the mark with distilled water. -4 5-ml. aliquot of Solution D is desirable for samples containing from 5 to 50% silicon. The solution is mixed well, transferred to a colorimeter tube, and allowed to stand for 30 minutes before the percentage light transmittance is read with a 400-mp light filter. Reference to the standard curve gives the micrograms of silicon per 50 ml. The percentage of silicon in the sample is (micrograms of silicon per 50 ml.) X (O.O5/ml. in aliquot X weight of sample). ALUMINURI DETERMINATION
Reagents. Hydrochloric acid, 1 to I; ammonium hydroxide; ammonium hydroxide, 1 to 4; ammonium chloride, 170 solution. Sodium hydroxide solution, approximately 25 grams of sodium hydroxide dissolved in 100 ml. of distilled water (this solution is made up fresh for each set of determinations and is used while hot). Topreparethebuffersolution, pH 4.2, approximately60ml. of glacial acetic acid are diluted to about 900 ml. with distilled water, a 100-ml. portion of 1070 sodium hydroxide solution is added, and the pH, as measured on the glass electrode is adjusted with small increments of 10% sodium hydroxide. For the aluminon reagent, exactly0.200 gram of aluminon, synthesized by the method of Smith et al. ( 6 ) ,is dissolved in 100 ml. of pH 4.2 buffer solution, and the volume is adjusted to 500 ml. with distilled water. Standard aluminum solution, stock solution containing 5 micrograms of aluminum per ml. of solution. iiliquots (1, 2, 3, 4, 5, 6, 7 , and 8 ml.) of this solution are taken for the standard curve, and the color is developed as described below, the aliquots of standard solution being substituted for the Solution F aliquot. Procedure. To the 50-ml. portion of Solution C (the supernatant solution from the silica centrifugation) in a 60-ml. pointed centrifuge tube are added 3 drops of bromocresol purple. The tube is brought to a boiling water bath temperature and concentrated ammonium hydroxide is then added until the first traces
628
ANALYTICAL CHEMISTRY
of precipitate start to forin, the solution being agitated constantly with the air-jet stirrer. Dilute ammonium hydroxide (1 to 4) is then added dropwise until the indicator turns from yellow to purple. The tube is placed in a hot water bath for 5 minutes and cooled, and the volume is adjusted to exactly 60 ml. If the indicator has changed back to yellow, dilute ammonium hydroxide is added until the purple color reappears. The suspension is mixed thoroughly with the air-jet stirrer, then centrifuged for 5 minutes a t 1800 r.p.m. The supernatant solution contains calcium, magnesium, and manganese, while the precipitate contains the hydroxy oxides of iron, titanium, and aluminum. A 50-ml. aliquot of the supernatant liquid is pipetted with a Lowy pipet into a 125-ml. conical flask, care being taken that the precipitate is not disturbed. This is Solution E, used for the determination of calcium plus magnesium. Approximately 50 ml. of a 1% ammonium chloride solution are then added to the tube, and the precipitate is thoroughly mixed with the air-jet stirrer. The suspension is again centrifuged a t 1800 r.p.m. for 5 minutes, after which the supernatant liquid is drawn off by gentle suction and discarded. The precipitate is dissolved by the addition of 3 ml. of 1 to 1 hydrochloric acid, and the suspension is stirred until solution is complete, Next, 10 ml. of hot sodium hydroxide solution are added with stirring, and the suspension is placed in a hot water bath for 5 minutes and then allowed to cool. I t is then diluted to exactly 50 ml. with distilled water and mixed with the air-jet stirrer. The stirrer is allowed to drain completely, but the adhering solution is not washed back into the tube. The suspension is centrifuged for 5 minutes a t 1800 r.p.m. 4 5-ml. aliquot of the supernatant solution is transferred to a 100-ml. volumetric flask, and 3 ml. of 1 to 1 hydro,chlorjc,acid are added. The solution is diluted to the mark with distilled water and mixed. This is Solution F, used for the determination of aluminum.
Table I.
Analyses of Bureau of Standards Samples
[AIlcrochemical system compared t o those obtained by h-ational Bureau of Standards (standard) I ArFillaceous Limestone Eiements Plastic Clay, No. Y8 x o 1.4 Determined Found Standard Found Standard
SiOz AlrOa Fez08 TiOz CaO
MgO 5820 KZO
59.6 25.7 2.07 1.44
0.21 0.69 0.24 3.26
59 1 25 5 2.05 1.43 0 21 0.72 0.28 3.17
14.2 4 07 1.60 0.19 41.6 2.23 0 31 0.76
14.1 4 16 1.63 0.16 41.3 2.19 0.39 0.71
+410-ml. portion of pH 4.2 buffer solution is placed in a 50-ml. volumetric flask, and the volume of the solution is adjusted to about 30 ml. Exactly 10 ml. of 0.04% aluminon reagent are added, and the flask is swirled to mix the solution. Finally, a Solution F aliquot containing 5 to 35 micrograms of an aluminum is added, and the solution is immediately diluted to the mark and the contents mixed well. (95-ml. aliquot is suitable for samples containing 1 to 8% aluminum.) The solution is transferred to a colorimeter tube, and the per cent light transmittance is read 25 minutes after the addition of the sample solution. -1520-mk light filter is used. Reference to the standard curve gives the concentration of aluminum in solution. The percentage of aluminum in the sample is (micrograms of aluminum per 50 ml.1 ml. in aliquot) X (O.lZ/weight of sample). DETERMINATION OF CALCIUM PLUS MAGNESIUM
Reagents. Ammonium hydroxide; 0.0100 N Versenate solution (prepared as described in the calcium determination). Eriochrome black T (Eastman) indicator solution, a 0.5-gram portion of Eriochrome black T mixed with 4.5 grams of hydroxylamine hydrochloride and dissolved in 100 ml. of methanol. Procedure. To the 50-ml. of Solution E contained in the 125ml, conical flask are added 5 ml. of concentrated ammonium hydroxide and 5 drops of Eriochrome black T indicator solution. The solution is then titrated with 0.0100 N Versenate solution to a bright blue end point. (Copper, cobalt, or nickel, if present, will interfere with the end point, and sodium cyanide must be added to complex these ions.) This titration is a measure of the total calcium plus magnesium in the solution. To obtain the percentage of magnesium in the sample, the calcium equivalent must first be subtracted., Thus (ml. of Versenate) X 0.0144/(weight of sample) = (milliequivalents of calcium plus magnesium per gram of sample). From the calcium determination (ml. of
Versenate) X 0.024/(weight of sample) = (milliequivalents of calcium per gram of sample). Therefore (milliequivalents of calcium plus magnesium per gram) - (milliequivalents of calcium per gram) = (milliequivalents of magnesium per gram of sample), and (milliequivalents of magnesium per gram) X 1.216 = percentage of magnesium in the sample. ACCURACY
An analysis of the errors involved in this analytical system shows that the over-all accuracy should be within approximately +2QJ,. There is an inherent error of approximately ~ 4 ~ 1in% volved in colorimeter readings. The error involved in making volume measurements in the pointed centrifuge tubes is also approximately +l%. The flame photometer can be assumed to be accurate within &2y0 in favorable concentration ranges, and the Versenate titrations involve absolute errors of approximately + O . l ml., which is equivalent to approximately 0.03% of calcium or magnesium. Thus, the maximum relative inherent error for the individual determinations (except in the very low concentration ranges) will be approximately as follows: silicon, & l % ; iron, + l % ; titanium, i l % ; calcium, aluminum, *3%; i l y o plus titration error; magnesium, 5 3 % plus double the titration error; sodium, +2%; and potassium, 5 2 % . Thus, i t is apparent that the maximum inherent error for the complete analysip should he approximately +Z%, RESULTS
Two samples obtained from the National Bureau of Standards were analyzed to check the accuracy of the new system against standard gravimetric procedures. The two samples, plastic clay and argillaceous limestone, were picked because they exhibited rather extreme variations in their compositions, The results of these analyses are presented in Table I. The values presented for the new system are averages of three determinations. The precision of the individual determinations was within the limits discussed in the section of accuracy. This system has been applied successfully in this laboratory to the analysis of a variety of materials including soil, clays, t r o p ical soils containing up to 70% iron oxide and 35% titanium oxide, and various relatively pure minerals exhibiting a vast range in chemical composition. In the separation of aluminum from iron, there was noticeable coprecipitation of aluminum on large amounts of hydroxy oxide precipitates of iron and titanium, resulting in low aluminum results. This situation could probably be corrected by converting the precipitated iron oxide t o ferrous sulfide by the method presented by Snell and Snell(7). However, this coprecipitation is not of great significance if the iron oxide content of the sample is less than 15%. LITERATURE CITED
(1) Dienert, F., and Waldenbulcke, F., Compt. rend.. 176, 1475-50 (1923). (2) Gysling, H., and Schwarzenbach,G., Helv. Chim. Acta, 32, 1454504 (1949).
(3) Hedin, R., “Colorimetric hfethods for Rapid Analysis of Silicate Materials,” Swedish Cement and Concrete Research Institute, Royal Institute of Technology, Stockholm, 1947. (4) Schwaraenbach, G., and Biedermann, W , , Hell;. Chim. Acta., 31, 678-87 (1948). ( 5 ) Schwaraenbach,G., and Gysling, H., Ibid., 32, 1314-25 (1949). (6) Smith, W. H., Sager, E. E., and Siewers, I. J., ANAL. CHEM., 21, 1334-8 (1949). (7) Snell, F. D., and Snell, C. T., “Colorimetric Methods of Analysis,” 1-01. 11, 3rd ed., Iiew Tork, D. Van Nostrand Co., 1949, (8) Killard, H. H., and Cake, IT. E., J . Am. Chem. Soc., 42,2208-12 (1920). (9) Yoe, J. H., and Armstrong, A. R., ANAL. CHEM., 19, 100-2 (1947).
(10)Yoe, J. H., and Jones, A. L., TND. ENG.CHEM., ANAL.ED., 16, 111-15 (1944).
RECEIVED for review August 11, 1952. Accepted December 29, 1952.
