DISCUSSION A N D RESULTS
Many of the extraction procedures currently available are based on solute differences in partition behavior a t equilibrium. I n these procedures, short separation times usually produce loiver separation factors than longer times, because equilibrium is not attained a t the short times. However, the separation of 110 from RB by solvent sublation is considerably better a t shorter times as seen in Table 11. Herein niay lie the potential of solvent sublation. Because of the kinetic processes occurring in sublation, time may be used as an effective parameter in the development of selective separations. It is of interest to note that solvent sublation differs from the more familiar liquid-liquid extraction technique, in which extensive mixing of the two immiscible phases occurs in the shaking
process. Mixtures of 310 and RB n i t h H D T have been studied under similar conditions as the solvent sublation experiments. An emulsion in both phases was observed after the shaking period; this emulsion broke up only after a 20-hour waiting period. hnalysis revealed separation factors on the order of 1.7, with evtensive eytraction of both AI0 and RB into the octanol phase. Thu;, it appears that solvent sublation allovs the use of surface active agents for selective extraction, iyhereas liquid-liquid extraction is not wited for these agents because of emulsion formation. Further, solvent sublation niay be more selective than liquidliquid extraction in certain cases, especially where differences in rates of extraction can occur. S o detailed interpretation of the niechanisiii of extraction for the two
dyes by solvent sublation is posqible with the limited amount of data presented here. Thus, a comparison of the tlvo techniques in ternir of mechanism cannot be given a t this time. Work is continuing in elucidating the mechanism of estraction of A10 and RB, and application of solvent sublation to other mixtures. The reqult, of these studies will be reported subsequently. LITERATURE CITED
(1) Sebba, F., “Ion Flotation,” Chap. X,
Elsevier, Sew York, 1962.
ALEGRIAB. CAR.4GA-Y
BARRYL. K ~ R G E R Department of Chemistry Northeavtern University Boston, PIIasv. WORKsupported by the Food Division of the U. S. Army Natick Laboratory under Contract No. DA 19-129-AMC-302(S).
N e w Direct Spectrophotometric Determination of Aluminum in Steel, Spelter, and Iron Ores SIR: Direct photometric methods employing Eriochrome Cyanine R dye for the determination of aluminum in iron ores and spelters have been described ( 1 , 2 ) . The application of the method to the routine analysis of iron ore and spelters has been satisfactory for both accuracy and the speed of analysis. Improvement in the direct photometric method for the determination of aluminum in steels, however, is desirable. The compensating iron blank requires extreme care in its preparation, as does the background color formation in the aliquot of the steel sample if accurate results are to be obtained ( 2 ) . The present work describes a new aluminum complex and its application to the direct spectrophotometric determination of even trace amounts of aluminum in alloy and carbon steels. The need for a compensating iron blank is eliminated and improved sensitivity and accuracy are realized. The extremely stable colored compound of aluminum is formed by combining a heterocyclic fatty acid amine with the aluminum-Eriochrome Cyanine R complex. The aluminum complex obeys the Beer-Lambert law and has the characteristic stability usually found in amine coordination compounds. Besides permitting the complete blocking of the iron color by the addition of a sulfite, a separation of some overlapping interference peaks takes place by a shift of the aluminum absorption peak toward the infrared. The method may be applied to the determination of aluminum in a wide variety of metallurgical products including stainless steels and alloys. 654
ANALYTICAL CHEMISTRY
EXPERIMENTAL
Apparatus and Reagents. A Beckman DU spectrophotometer was used. One gram of Eriochronie Cyanine R (General Dyestuff Corp., Melrose Park, Ill.) was dissolved in water and diluted to 1 liter. The ammonium acetate buffer solution m s prepared by dissolving 320 grams of aluminum-free ammonium acetate in nater, adding 25 grams of sodium sulfite, dissolving, and diluting to 1 liter. The p H was adjusted to 7.6 by the addition of ammonium hydroxide or acetic acid as needed. The ammonium acetate-polycyclic ketoamine working buffer solution was prepared by adding 0.7 gram of polycyclic ketoamine (Amcheni Products, Inc., Ambler, Pa.) to 100 nil. of ammonium acetate buffer. The dissolving acid solution was 3-17 sulfuric acid-3S nitric acid 4 : 1 (v./v.). Procedure for Steels. XCIDSOLUBLEAALUhIINUJI. TO a 0.5000gram sample in a 300-ml. Erlenmeyer flask, add 50 ml. of dissolving acid and place in solution on a hot plate. Boil out t h e oxides of nitrogen and oxidize the carbon by adding a drop or tw-o of saturated potassium permanganate solution in excess. Reduce t h e permanganate by t h e dropwise addition of 10% sodium nitrite, or saturated sulfurous acid, and boil briefly. Cool and dilute to 250 ml. and mix. Dilute 40 ml. of 3 5 sulfuric acid to 250 ml. to serve as a blank. Place 2 ml. of sample and 2 ml. of blank solution into two separate 25-inl. volumetric flasks. To each, add 2 ml. of 0.1% Eriochrome Cyanine R solution and allolv to stand for a few minutes or warm briefly in hot water (66-100” C.).
Add 2 ml. of 2% mercaptoacetic acid and 2 nil. of cyclo-ketoamine, qodiuni sulfite, ammonium acetate buffer. Allow to react for about 2 minutes and then dilute to 25 nil. and mix. Heat in boiling water for about 30 seconds and allow to stand for about 7 minutes at room temperature. Cool and obtain the absorbance of the sample against the blank a t 595 nip in a 1-cni. cell. From a previously prepared calibration curve, obtain the per cent aluminum present. TOTAL ;ILUUINUM. Nethod .1. Dissolve the sample as in acid-soluble aluminum. Filter on Whatman 42 filter paper, wash, and ignite the residue. Fuse the acid-insoluble residue with potassium b i d f a t e in a platinum crucible. Cool and dissolve the melt in the filtrate. Determine aluminum as in the acid-soluble method. Method B. Dissolve the sample as above and take to fumes of sulfuric acid. Cool and take up the salts in 20 ml. of 1:1 hydrochloric acid. Proceed as in acid-soluble aluminum. ALLOY STEELS. T o a 0.4000-gram sample in a 300-ml. Eilenmeyer flask add 5 nil. of 30% hydrogen peroxide, then cautiously add sufficient hydrochloric acid to dissolve the sample. When in solution, add 10 ml. of nitric acid and 15 ml. of perchloric acid. Take to strong fumes of perchloric acid and remove the chromium with anhydrous hydrogen chloride. Cool the solution in ice water, dilute to 50 n i l , and add a sufficient evceqs of cold 6% cupferron solution to precipitate all of the vanadium, titanium, and zirconium. Illow to react a few minutes and dilute to 100 nil. Filter a portion through a Whatman 40 paper, divarding the first portion of filtrate. From the second portion, pipet a 1.0-ml.
Figure 1. Comparison of absorption spectra of alurninumand without (-) polycyclic Eriochrorne Cyanine R with (-) ketoamine
--
1 -cm. cell 0.2 pg. of AI p e r ml.
aliquot into a 25-ml. volumetric flask. Proceed as for acid-soluble aluminum. Preparation of a Calibration Curve. Dissolve a U.S. Bureau of Standards steel, such as 55e, as described in t h e procedure. Transfer 0, 1, 2, 3, 4, 5, 6 ml. of standard aluminum solution (1 ml. = 1.0 pg.) to 25-m1. volumetric flasks, respectively. Add 2 ml. of the standard steel solution to each flask. Develop the aluminum color as described in the procedure. Measure the absorbance of each sample against the pure steel sample containing no added aluminum. Procedure for Spelters. Weigh a 0.2000-gram sample and proceed as for s t e d s except -omit t h e oxidation step. Procedure for Iron Ores. T o a 0.1000-gram sample, in a platinum crucible, add three drops of 1: 3 sulfuric acid and 1 to 2 ml. of hydrofluoric acid. Take to dryness on a hot plate and ignite briefly. -4dd 2 grams of potassium bisulfate and fuse the residue. Cool and dissolve the melt in 80 nil. of 31%' dissolving acid. Cool and dilute to 250 ml. Pipet a 1-ml. aliquot into a 25-ml. volumetric flask and proceed as for steels. Measure the absorbance against the reagent blank. DISCUSSION
Effect of Sodium Sulfite, Ethanol, a n d Polycyclic Amines. T h e effects of sodium sulfite and other reducing compounds upon t h e Eriochrome aluminum complex have been reported ( I ) . It was found, however, that if ethanol were added with the dye almost all of the iron coniplex was destroyed, yet permitting partial formation of a n aluminum complex. Stabilizers which were more effective than ethanol for the aluminum complex were found in the tertiary, heterocyclic, high-molecularweight derivatives of rosin amines or
those of lauric, oleic, and stearic acids. Only the rosin derivative has been employed in these studies, but a test of the other amine derivatives indicates that substitutions may be made in the proposed method. For example, the lauric acid derivative, dissolved in dimethylsulfoxide and then combined in the buffer, produced an identical calibration curve as that obtained with the rosin derivative at a slightly higher concentration of the amine. Color formation is instant on addition of a single drop of the rosin amine in the presence of buffer and sodium sulfite, resulting in a new aluminum complex of increased sensitivity and stability. The molar absorptivity of the new complex is 105,or on the Sandell scale it is 3.52 X f i g . of Al/sq. cm. The color is stable for five or more hours. Figure 2 shows the effect of polycyclic ketoamine concentration on absorbance of the aluminum complex. Triethanol amine and other lower weight amines stabilize but do not enhance the sensitivity of the aluminum complex to the same extent as the heterocyclic amines, Effect of pH. T h e optimum p H for the color formation is between 5.3 and 5.6. T h e color formation does not vary greatly between these values b u t drops off rapidly a t higher values.
Table 1.
T h e choice of dissolving acid was made on the basis of the high boiling point of sulfuric acid in order to maintain an even p H value. Other acids may be employed, such as perchloric, hydrochloric, sulfuric, and nitric, as long as the final solution is about 0.5 to 0.6N. Ammonium acetate buffer is' preferred over sodium acetate to adjust the pH because the rosin amine derivative is not completely soluble in sodium acetate. Absorption Spectra. Figure 1 shows t h e absorption spectra of t h e aluminum complex with and without t h e addition of the rosin amine derivative. The position of the absorption peak in relation to wavelength is a function of the temperature and time of complex formation, type and concentration of t h e masking compound, and the dissolving acid. A 5% solution of ascorbic acid may be substituted for the mercaptoacetic acid for masking the iron. The ascorbic acid addition results in a more rapid complex formation and a wider shift of the absorption peak toward the longer wavelength. Ascorbic acid is less stable, however, than mercaptoacetic acid and must be prepared fresh a t least daily, while mercaptoacetic acid remains stable for a year or longer. Effect of Temperature. Immersion of the sample flask in hot water for 1 minute reduces the time for the color formation from 20 minutes, a t room temperature, to about 5 minutes. T h e sample may also be diluted with hot water, then cooled, and made u p to volume. After the complex has formed, it is not influenced by small changes in room temperature and i t remains stable for long periods. Analytical Reagents. Characteristics of Eriochrome Cyanine R have been previously investigated (I). Xitration of the dye to improve the stability of the aluminum complex was unnecessary in the presence of polycyclic ketoamine. The most serious difficulty encountered with ammonium acetate was from aluminum contamination as a result of contact, in a deliquescent state, with the glass container. When an anhydrous supply is obtained i t should be repacked in polyethylene to prevent aluminum contamination. The buffer solution should be stored in a plastic container. Four separate commercial batches of
Analysis of Standard and Miscellaneous Steels
170A Titanium bearing 106 Chromium, molybdenum, aluminum 55e Open hearth iron 339 Stainless 65d Basic electric
Present 0,046 1.06 0,002 0.080
0,059
Aluminum Found 0,050 1.06
0.003 0.078 0.060
Deviation +0.004 0,000 +O.OOl
-0.002 $0.001
VOL. 38, NO. 4, APRIL 1966
655
Table IV.
Repeat Analysis
Alu-
Sample 65d
Mean DS-7
Mean Figure 2. complex
26
Table II. Analysis of Standard and Miscellaneous Spelters
Present A B
C D
Aluminum Found Deviation
0.28
0.280
0.000
0.15 0.13 0.13 0.12
0.152 0.130 0,132 0.120
$0.002 0.000 $0.002 0.000
Table 111.
0
from mean
Std. dev.
n nnn
0.060 0.000 -0.001 0.000 0.001 0,000 0.000 -0.001 0,000 +O,OOl 3 ~ 0 . 0 0 0 4 0.000710
0.059 0.060 0.061 0.060 0.060 0.059 0.060 0.061 0.060
0,280 0.280 0.282 0.278 0.283 0.278 0.279 0.280 0.280 10.00128 0,00188
sulfite before the addition of water. The buffer so prepared remains clear and stable for a month or longer. The sodium sulfite may also be combined with the mercaptoacetic acid and the ketoamine with the ammonium acetate, as was done in preparing Figure 2. Interfering Elements. The number of interfering elements is reduced when comparison is made with the previous method (2). Cobalt, for instance, reacts with Eriochrome Cyanine R but does not shift its absorption peak with that of aluminum, although it interferes a t 535 mp. Chromium does not produce a negative interference as formerly, but forms a weak complex with a maximum at 560 mp. A 1% chromium content produces a positive error of 0.002% aluminum. Titanium, vanadium, and zirconium produce positive errors equal to 0.05’% aluminum for 1% of these elements. The interferences are conveniently removed with cupferron as described in the method for alloys.
Mean
0.00 0.00 -0.01 0.00 $0.01 0.00 0.00 $0.01 -0.01 zk0,0044
1.02 1.02 1.01 1.02 1.03 1.02 1.02 1.03 1.01 1.02
ANALYTICAL CHEMISTRY
0.00758
RESULTS
Table I shows the results obtained with the method on various steels. Tables I1 and I11 show the results obtained on spelters and iron ores, respectively. Excellent agreement with established values has been obtained. Table IV shows the results of repeat analyses of steels, spelter, and iron ore.
LITERATURE CITED
(1) Hill, U.T., ANAL.CHEW 28, p. 1419 (1956). (2) Ibid., 31, 429 (1959).
Analysis of Standard and Miscellaneous Ores
NBS Sample 26 Magnetite NBS Sample 29 Magnetite NBS Sample 29A Sibley NBS Sample 27B
656
%
n ntin
Effect of polycyclic ketoamine concentration on absorbance of aluminum
polycyclic ketoamine, manufactured a t intervals of several years, were tested in the procedure. No deviation in the results was noted with these separate samples. To ensure a uniform reagent, polycyclic ketoamine is being made available in a reagent grade. Slight turbidity, which a t times may develop when polycyclic ketoamine is added to an aqueous buffer solution, may be avoided if the ketoamine is added to the solid ammonium acetate and sodium
DS-7 Sample
minum, Deviation
UNOT. HILL
Present
Aluminum Found
Deviation
1.02 1.91 0.46 0.59
1.02 1.92 0.46 0.59
0.00 +0.01 0.00 0.00
Inland Steel Co. East Chicago, Ind. PRESENTED at the Sixteenth Annual MidAmerica Spectroscopy Symposium, Chicago, Ill., June 1965.
