(11) Habgood, H. W., Harris, W. E., Ibid., 32,450 (1960); 34,882 (1962). (12) Huguet, M., JournBes Intern. de SBparation ImmBdiate et de Chromatographie, p. 69, G.A.M.S., Paris, 1962. (13) Kaiser, R., “Chromatographie in
der Gas phase,” Bibliographisches Inatitut, A.G. Mannheim, 1962, Part 111, Tabellen. KID. 78-115: “Gas Phase Chromatography 111,” Butterwortha, London, 1963. (14) Kovats, E., Helv. Chim. Acta 41, 1915 (1958). (15) Landault, C., Guiochon, Chromatog. 9 , 133 (1962).
G., J.
(16) Loewenguth, J. C., Vth Petroleum
World Congress, Preprint, Frankfurt,
1963. (17) Martin, A. J. P., Biochem. SOC. Symposia 8 , 4 (1949). (18) Merritt, C., Walsh, J. T., ANAL. CHEM.34,903 (1962). (19) Ibid., 908. (20) Pitkettly, R. C., Ibid., 30, 1209 11958). ( 2 i ) Poher, P. E., Deal, C. H., Stross, F. H., J . Am. Chem. SOC. 78, 2999 (1956). (22) Strickler, H., Kovats, E., J . Chromatog. 8 , 289 (1962).
(23) Swoboda, P. A. T., “Gas Chromatography 1962,” p. 273, Van Sway, M . ed., Butterworths, London, 1963. (24) Wehrli, A,, Kovats, E., Helv. Chim. Acta 42, 2726 (1959). (25) Zulaica, J., Guiochon, G., Bull. SOC. Chim. France 1963, 1246. (26,) Zulaica, J., Guiochon, G., Unpublished data, Paris, France, 1963.
GEORGES GUIOCHON Laboratoire du Professeur L. Jacque Ecole Polytechnique Paris ( 5 e ) , France
Determination of Small Amounts of Manganese by Oxidation of 8-Aminoquinoline, Extraction, and Spectrophotometry SIR: Oxidations of the organic reagents o-tolidine (S), benzidine (9), tetramethyldiaminodiph e n y 1m e t h a n e (8),and tetramethyldiaminotriphenylmethane (4)with permanganate have been used as the basis for the determination of manganese. Although these are sensitive methods, the colored products are unstable. 8-Aminoquinoline is oxidized to a highly colored oxidation product with permanganate and the color formed is stable. 8-Aminoquinoline has been used for the gravimetric determination of copper ( I ) , and for the spectrophotometric determination of palladium (6) and iron (6). EXPERIMENTAL
Reagents and Apparatus. &AMINO-
QUINOLINE SOLUTION. 8-Aminoquino-
line, 50 mg., was dissolved in 1 ml. of concentrated sulfuric acid and diluted to 100 ml. with distilled water.
STANDARDPERMANGANATE SOLUA stock solution of Mn(VI1)
TIONS.
was prepared by dissolving 0.10 mole of reagent grade K M n 0 4 in distilled water and diluting to 1 liter. The stock solution was standardized against primary qtandard grade sodium oxalate (7‘). All other Mn(VI1) snlutions were prepared by appropriate dilution of the standardized stock solution. STANDARD M ~ ( 1 1 SOLUTIONS. ) Stock solutions of hln(I1) were prepared by
dissolving reagent grade M n S 0 4 in dilute sulfuric acid. The solutions were standardized spectrophotometrically after oxidation with bismuthate ( 2 ) . SOLVENTEXTRACTION SYSTEM.The benzyl alcohol-chloraform solution used for the extraction was prepared by mixing equal volumes of benzyl alcohol and chloroform. All solutions used in the interference studies were prepared by dissolving reagent grade salts in distilled water. Dilute HnSOa was added to prevent hydrolysis. The apparatus is the same as that described in a previous publication (6). Procedure. To a 50 ml. sample containing 1 to 30 pg. of h h ( I I ) , add 3 ml. of concentrated sulfuric acid and 0.2 gram of NaBiOs. Stir for 1 minute and filter through a 60-ml. sintered glass funnel into a 125-ml. filter flask. Wash the excess solid XaBiOs on the filter with two 10-ml, portions of 1M HzS04. Transfer the filtrate quantitatively to a 150-ml. beaker, add 3 ml. of the 8-aminoquinoline reagent solution, and then add solid NaOH until the p H is 11.0 or greater. Stir the solution for several minutes and adjust the p H to the interval 1.45 to 1.95 with 1 to 1 HzS04. Transfer the solution to a 125-ml. short stem separatory funnel, add 5.00 ml. of the benzyl alcoholchloroform solution and extract, shaking occasionally over a period of several minutes. Fill a 1-cm. absorption cell with the organic phase and measure the absorbance a t 550 mp against a blank. The hlank is prepared by treating 50 ml. of distilled water in the same manner as the manganese sample.
DISCUSSION A N D RESULTS
When Mn(VI1) reacts with 8-aminoquinoline, a red-colored oxidation product is formed which is soluble both in water and in a mixture of benzyl alcohol and chloroform. The absorption curve for the benzyl alcoholchloroform extract of the oxidation product is the same as that obtained when the reagent is oxidized by ferric iron (6), with maxima at 517 and 550 mp. The oxidation of 8-aminoquinoline with Mn(VI1) is sensitive to pH. The effect of the initial p H on the reaction was studied and the results are given in Figure 1. The solutions were prepared by adding a 50-ml. sample containing 14.7 pg. of Mn(VI1) to 3 ml. of the reagent solution. The pH of each of these solutions was adjusted to the desired value before they were mixed. After being mixed, the solutions were stirred for several minutes. The p H was then adjusted to 1.60, the solution was extracted with 5 ml. of the benzyl alcohol-chloroform solvent, and the absorbance was measured a t 550 mp. Oxidation is incomplete a t low p H values. If the reaction mixture is first made basic a t a p H of 10.8 or greater and then acidified to a p H of 1.60, the reaction proceeds quantitatively. Two 50-ml. samples that contained 44.1 pg. of Mn(VI1) were made basic to a p H of 11.O. They were added to 3 ml. of the reagent solution which was previously
Figure 2. Absorb-
;
0 4
Figure 1.
b
PH
A
Mn(VII) vs. pH
,b
_J I2
Effect of pH on oxidation
14.7 fig. of Mn(VII) used for oxidation
1674
ANALYTICAL CHEMISTRY
aool
I
2
I
3
I
4
I
5
I
6
I
7
I
8
I
9
I
IO
In
II
12
08
OE
h E 0
-
2 w D
4
x
j 4
02
00 000
I 025
I
Aqueous solutions used to measure absorbance
adjusted bo a pH of 11.0, and a browncolored precipitat,e of MnOz was observed. Solution A was filtered and then made acidic to a 1111 of 1.6. The pH of solution I3 was adjusted to 1.6 without filtering, and the absorbance of both solutions was measured a t 550 mp. The absorbance of solution was only 61% that of solution 13. In another study t'he extent of reaction was evaluat,ed by measuring the amount, of permanganate remaining in the> aqueous phase after reaction and extraction of the osidation product. To do this, the concentration of the Mn(VI1) had to be increased to obtain sufficiently high absorbance readings. Solutions were prepared by mising 25ml. samples of Mn(T'I1) that contained 275 pg. of Mn(VI1) with 3 i d . of the reagent solution. (The pH of each h(J1Ution was adjusted to the desired value before mising.) The pH was adjusted to 1.6, the solutions were estracted with 5 ml. of the benzyl alcohol-chloroform mixture, and the organic phase was discarded. The aqueous phase wss diluted to 50 ml. and the absorbance was measured a t 530 mp, the absorbance nirisimum for the permanganat8e ion. Results are shown in Figure 2. h 25ml. solution containing 275 pg. of LIn(VII) diluted to 50 ml. had an absorbance of 0.184 a t 530 mp. The results confirmed t'hose obtained by the measurenient of the osidation product formation above, indicating a comiilete rcwtion at :L1)H of 10.7. From the results of the preceding studies it can be assumed that, when initial pH values lower than 10.7 are uwd, the oxidation of 8-aminoquinoline A\
is incomplet'e and that the formation of Mn(IV) from Mn(V11) is the incomplete step in t'he overall reduction process. The number of electrons involved in the osidat'ion can be determined by finding the number of moles of permanganate needed t o react with a given amount of reagent. The results of t,his st'udy are given in Figure 3-1.00 x mole of 8-aminoquinoline required 4.0 X mole of permanganate for complete osidation. Since the reduction of permanganate proceeds by the reaction MnOa- += Mn+2, a t'wo-electron change in the oxidation of the reagent is indicated. Results of a study to determine the effect of pH on the extraction are given in Figure 4. Solutions were prepared by reacting a 50-ni1. sample that contained 14.7 fig. of Mn(VI1) a t pH 11.0 with 3 nil. of the reagent solution a t a pH of 11.0 and then adjust'ing the pH to t'he value indicated. Optimum pH range fnr the estraction is 1.45 to 1.95. At higher and lower acidities, the color of t8he osidation product is dwtroyed. If the solutions are osidized and then maintained a t a pH within this interval, the color is extremely stable with respect to time. Solutions measured t,wo weeks after extraction showed no change in absorbance. In contrast to the osidation of 8aniinoquinoline with Fe(I1I) ( 6 ) there is no salt effect' on the oxidation with hIn(VI1). Solutions of 8-aminoquinoline osidized in 1.5M S O 3 - ! SO4-*, and C104- media gave the same absorbance as one osidized in the absence of a salt. Chloride, however, did have an rffect on t'he oxidation. Ai1JI C1solution decreased the extent of osidation with 14.7 nig. of Mn(VI1) by 20%. The effect of chloride is probably caused by the reduction of permanganate by chloride, which decreases the amount of permanganat'e available for the oxidation of 8-nminoquinoline. KO salt effect on the estraction was observed. The selection of an agent to oxidize Mn(I1) to IZIn(S'I1) is important because it is necwsarg- to remove completely the escess osidizing agent before t'he addition of 8-aminoquinoline. If the escess is not removed, it too will osidize the 8-aminoquinoline, yielding a high result for manganese. 130th periodate and persulfate were investigated and found to be unsuitahle. For this osidation, bismuthate was wpcrior to all others tested. The bismuthate oxidation of manganese wa Cunningham and Coltrnan (2). ThfA escess bismuthate can be removed easily by filtration and the reaction proceeds a t room temperature in 1 minute. Because the oxidation of &aminoquinoline must be done in basic solution, tlic sulfuric acid solutions of perman-
t
OS
n
" PH
00 O
Figure 4.
L
Effect of pH on extraction
14.7 F g . of Mn(VII) used for oxidation
panate used in the bismuthate oxidation must be made ba.4c. Solid SaOH !vas used to prevent further increase in volume upon neut'ralization. I3eer's law is obeyed over the concentration range 0.6 t,o 27.5 p g . of Mn(I1) in a 5 0 4 . sample. The sensitivity of the method, expressed as the weight, of manganese corresponding to an absorbance of 0.010 a t 550 mp is 0.294 pg. This corresponds to a sensitivity of 0.0059 p.p.m. if a 50-ml. sample of unknown manganese solution is used, as suggested in the procedure. The osidation of 8-aminoquinuline with permanganate proceeds rapidly at, room temperature even at extremely low
Table I limiting Concentration of Other Oxidizing Agents
Limiting conrn.,
Interference Fe( I11j I-(1.) Ce(IT)
D.D.m. .
25 4 I72 31 8
Cr(I'1)
Table II.
Ion added
I
468
Analysis of Synthetic Samples
Weight
Mn found, W t . , pg.
I
3In(II) Cu(I1) .%l(III) Zn(I1) Ca(I1) Mg(I1) Mn(I1)
T(1.)
Ce(IT*)
Cr(1.I) Fc(II1)
LIn(1t)
FelIII) Cr(Y1) Zn(I1)
Rlg(I1) Ca(1I)
11 1 1 1
0 pg.
12 mg.
07 mg. 2 3 mg 9 84 nig 10 2 mg
I1 16 5 p g . 1 04 mg. 2 15 mg.
201 pg.
6 41mg.
I11 2 7 5 pg. 128 p g . 101 p g 114 pg.
a,
11 2 11 0 10
I)
11 0
1. 16 i 2. 16 6 3. 16 8 Av. 16 7
1. 2.
3
2 55 ing. 2 46 nig.
VOL. 36. NO.
1. 2 3. .4v.
2 00 2 63 2 hG
Av. L' 79
JULY 1964
1675
concentrations. This is not true of other oxidizing agents such as Fe(III), Cr(VI), V(V),and Ce(1V). I n addition to the great effect of temperature and concentration on the rate of oxidation of 8-aminoquinoline when these agents are used, there are also different pH requirements for the oxidation. All of these factors tend to decrease the interference of these materials in the manganese determination. The results of an interference study to determine the limiting concentration of other oxidizing agents in the manganese determination are given in Table I. Limiting concentration is defined here as the concentration needed to cause an error of 0.010 in the measured absorbance of manganese. I n addition to the interfering ions, all solutions contained 11.O pg. of manganese.
Because the method developed will be particularly useful for small amounts of manganese it should find great utility for manganese trace analysis in water. Therefore, in addition to other oxidizing agents, many ions present in natural waters were also checked as interferences. The addition of 25 p.p.m. of C U + ~ , Zn+2, or K + to 0.22 p.p.m. Mn(I1) caused no error. Results were not affected by 200 p.p.m. C a f 2 and Mg +z* The method was checked by preparing three synthetic samples. Results are given in Table 11. LITERATURE CITED
( I ) Bankavskis, J., Latvijas PSR Zinatnu Akad. Vestis 9 , 115 (1955). ( 2 ) Cunningham, T. R., Coltman, R. W., Znd. Eng. Chem. 16, 58 (1924).
( 3 ) Forman, L., J . Am. W a t e r U'orks Assoc. 21, 1212 (1929). (. 4 .) Gates. E. M.. Ellis. G. H.. J . R i d . Chem. 168, 537' ( 1 9 4 i ) . 15) Gustin, V. K . , Sweet. T. R.. ANAL. CHEM.35, 44 (1963) ( 6 ) Zbid., p. 1395. ( 7 ) Kolthoff, J. RI., Sandell, E. B., "Textbook of Quantitative Inorganic Analysis," 3rd ed., p. 564, Rlacmillan, New York, 1952. ( 8 ) Prodinger, JT.) Jfikrochemie 36, 580 (1951). ( 9 ) Wiese, A. C., Johnson, B. C., J . B i d . Chem. 127, 203 (1939). 1-AUGHN K. GUSTIN'
THOMAS R. SWEET ,McPherson Chemical Laboratory The Ohio State University Columbus IO, Ohio Present address: Corning Glass Works, Corning, N. Y . m'ork taken in part from the Ph.D. dissertation of Vaughn K. Gustin, The Ohio State University, Columbus 10, Ohio, 1963.
Spectrophotometric Determination of Calcium in Milk Using 2,2 '-(E th a ned iy Iid ene d in itrilo)d iphe no I [Gl y oxa I Bis (2 - hy d roxya n iI)] SIR: Because of the wide distribution of calcium in nature, a great variety of procedures have been developed to measure it. Accompanying metals, however, often interfere with the chemical reactions involved in measuring calcium, so various means must be used to circumvent the interference. Calcium which plays a significant role in the physical and chemical properties of milk, must be measured accurately on a routine basis. Several reported methods are based on titration (ethylenedinitri1o)tetraacetate with (EDT.1) (1, 8, 9, If) and various indicators, but orthophosphate ions present in the milk may interfere with the end point ( 2 , 5, 6). Several methods (3, 7 , 10, fa) reported more recently are based on the color produced with 2,2'-(ethanediylidenedinitri1o)diphenol [glyoxal bis(2hydroxyanil) (GBH*\)], a Schiff base, where a red innercomplex salt is formed. Interference by other salts or ions can be blocked easily to give a specific test for calcium ( 3 ) . This paper presents a method for microestimation of calcium in milk that has the advantages of a colorimetric procedure. I t requires only a small sample and avoids the difficulty of determining an exact end point. These modifications are believed to have considerably improved Kerr's method (7') and yielded a convenient procedure adaptable to biological materials.
DB. Absorbance of the red calcium complex was measured with a Model D U Beckman spectrophotometer, using 1-cm. matched, square cells in each case. The pH values of buffers and reaction mixtures were checked with a pH meter. Reagents. STANDARD CALcIuhi SOLUTION. Dissolve 2.4972 grams of reagent grade calcium carbonate, dried a t 110" C., in concentrated hydrochloric acid. Make to 1 liter with deionized water. This solution contains 1.00 mg. of calcium per ml. ( 6 ) . GLYOXAL BlS (2-HYDROXYANIL) SOLUTION, 0.5%. Dissolve 0.5 gram of Fisher G-147 glyoxal bis(2-hydroxyanil) in 100 ml. of rediqtilled methanol ( 7 ) .
0.4
0.3
w U Z
2 0.Y
e
0
0.1
,
"
I
1
I
1
1
500
EXPERIMENTAL
Apparatus. The initial spectrum was determined with a Beckman recording spectrophotometer, Model 1676
ANALYTICAL CHEMISTRY
WAVELENGTH
1
1
550
1
1
I
1
1
600
(mp)
Figure I . Absorption spectrum of the reaction product of calcium and glyoxal bis2-hydroxyanil
TRIS BUFFER SOLUTIOKWITH PH 12.7. To 1 liter of 0.1M tris(hydroxymethylaminomet,hane) is added 100 ml. of 10% potassium hydroxide. AMMONIUMOXALATE-OXALICACID SOLUTIOK WITH PH 5 (4). Dissolve 27 grams of ammonium oxalate and 1.26 grams of crystalline oxalic acid and dilute to 2 liters with deionized water. All other materials were prepared directly from reagent-grade chemicals, except' that the methanol and et,hanol (95%) were freshly distilled before use. Procedure. h 2-ml. sample (concentrations of 0.40 to 4.00 mg. per ml. of calcium) is pipetted into 50-ml. round-bottomed centrifuge tube, and 10 ml. of ammonium oxalate-oxalic acid solution is added after the pH is adjusted to greater than 5 with 1Oyo potassium hydroxide, using 1 drop of Congo red as an indicator. The tubes are capped and allowed to stand for 1 hour. At the end of the timed period the tubes are placed in a centrifuge and spun for 15 minutes a t 3,000 r.p.m. The supernatant is carefully poured off. The precipitate is dissolved in I S HC1 (about 2 ml.). The contents are quantitatively transferred and diluted to give a calcium concentration within the desired range. For a skim milk sample, the precipitate is transferred to a 200-ml. volumetric flask and brought to ~ o l u m ewith deionized water. After a further 1 : l dilution, a 10-ml. sample is pipetted into a 125-ml. Erlenmeyer flask, and 5 ml. of buffer (tris/ KOH at pH 12.T) is added. The blank is 10 nil. of deionized water and 5 ml. of buffer (tris '%OH a t pH 12.7). The 0.5 ml. of color reagent, glyoxal bis(2hydrosyanil), and 10 ml. of ethanol are added in that order, with mixing, before proceeding t o the next flask. Absorbance was measured after a 10-minute color development period.