V O L U M E 26, NO. 1, J A N U A R Y 1 9 5 4 paper strip, so that the signal from the oxalic acid region should be less than in the absence of the film. By a similar argument, the signal from the dextrose region would not be affected to such an extent. DISCUSSION
The procedures described in the previous section have been shown to be capable of detecting a large assortment of chemical compounds and to be of some use in detecting the separated components on paper chromatograms. I n the procedure involving the comparison of dry paper strips, the measured signal indicating the presence of a sample on the strip appears to be due largely to the lower temperature of the surface containing the sample. A sizable contribution to the signal, however, is due to either a decrease in the transmission of the paper in the sample region or a lowering of the emissivity of the surface containing the a m p l e . No good method was found to distinguish between these latter possibilities. The relative insensitivity of the procedure in detecting sugars such as dextrose and maltose suggests that the chemical similarity of these substances with cellulose may give rise to similarities in the infrared emission and transmission of these substances. The measured signal in the comparison of oiled paper strips appears to be almost entirely due to the transmission of radiation from the heating mantle. Heat capacity effects were shown to be small and there is no reason to expect that the large increase in signal after application of the oil is due to larger differences in emissivities. This procedure is also relatively insensitive to sugars, again suggesting the influence of similarity in chemical structure between the sugars and cellulose. The greater sensitivity of this procedure together with the greater selectivity of transmission measurements over heat capacity measurements in identifying chemical compounds make this the preferred method. Chromatographic separations on paper usually involve material quantities in the order of several micrograms, whereas the quantities which can be successfully detected by this method are in the order of several milligrams. Now, as has been shown, chromatographic separations can be made with such large amounts on filter paper, but the relative movement of the bands must be widely different The major usefulness of paper chromatography on the other hand is its ability to deal with small samples and the trend of development for this technique appears to be in continually decreasing the sample size as soon as suitable detection methods become available. Thus, the present apparatus and technique cannot be considered as suitable for the detection of the majority of paper chromatograms.
195
By changing the arrangement of the apparatus 80 that the conditions for making transmission measurements are at an optimum, the detection limits of this technique might be extended down into a more useful range. Improvement in transmission measurements could be accomplished by increasing the infrared intensity emitted by the source after properly shielding the strips in some thermostated enclosure. It is possible that the application of other inert nonvolatile liquids to the strips would also improve their transmission properties. Because of the proximity of the noise level, a narrow band pass filter in the amplifier circuit would also be very desirable. Some indication has been shown that transmittance measurements coupled with the use of a negative filter can be used not only to detect a substance but also to identify it. Here again, however, more quantitative work would be necessary e0 establish this technique firmly. ACKNOWLEDGMENT
This research has been supported in part by the Abbott Fund of Northwestern University and in part by research grant RG2749 of the National Institutes of Health, PLiblic Health Service. LITERATURE CITED
Barr, E. S., itnd Chrisman, C. H., J . Chem. Phys., 8 , 5 1 (1940). ( 2 ) Carter, C. E., J . Am. Chem. Soc., 7 2 , 5 6 1 2 (1950). (3) Holiday, E. R., and Johnson, E. A,, Nature, 1 6 3 , 2 1 6 (1949). ( 4 ) Hotchkiss, R. D., J . B i d . Chem., 1 7 5 , 3 1 5 (1948). (5) Jamison, N. C., Kohler, T. R., andKoppius, 0. G . ,ANAL.CHEM., (1)
23, 553 (1951). (6) Jones, A. R., Ibid., 24, IO55 (1952). ( 7 ) Kivenson, G. J., J . Opt. SOC.Amer., 40, 112 (1949). (8) Lugg, J. W., and Overell, B. T., Nature, 160,87 (1947). (9) Martin, G. A., Instruments, 22, 1102 (1949). (IO) hliiller, R. H., and Wise, E. N., ANAL.CHEM.,2 3 , 2 0 7 (1951). (11) Paladine, A. C., and Leloir, L. F., Ibid., 24, 1024 (1952). (12) Pereira, A , , and Serra, J. rl., Science, 113,387 (1951). (13) Randall, H. hl., Fowler, R. G., Fuson, N.,and Dangle, J. R., “Infrared Determination of Organic Structures,” p. 105.
New York, D. Van Nostrand Co., 1949. (14) Rockland, L. B., Lieberman, J., and Dunn, M. S., ANAL.CHEM., 2 4 , 7 7 8 (1952). (15) Rowen, J. W., and Plyler, E. K., J . Research Natl. Bur. Standards, 44, 313 (1950). (16) Tennent, D. M., Whitla, J. B., and Florey, K., ANAL. CHEM., 2 3 , 1 7 4 8 (1951). RECEIVED for review February 9, 1953. Accepted October 1, 1953. Abstracted from the Ph.D. thesis of Donald R . Kdkwarf, Northwestern University, 1952.
Spect rophotometric Tit rations with Ethylenediaminetetraacetic Acid (II) Determination of Magnesium, Calcium, Zinc, Cadmium, Titanium, and Zirconium PHILIP B. SWEETSER’ and C L A R K E. BRICKER Department o f Chemistry, Princeton University, Princeton,
W
I D E use of ethylenediaminetetraacetic acid or its sodium salts (EDTA, Versenate, Sequestrene, or Complexone 111) as a volumetric agent has been possible because of the broad chelating power and stability of the Versenate chelates. Versenate forms a 1 to I chelate with B large number of di-, tri-, and, in some cases, tetravalent cations. By the proper use of buffers and additional complexing agents, the chelating power of the Versenate can be made very selective. This has been illustrated by Kinnunen and Merikanto in their procedure for the deter-
* Present addreese, Chemical Department, Experimental Station, E. I. du Pout de Nemoura L Co., h a . , Wilmington, Del.
N.1. mination of zinc in the presence of copper by the addition of cyanide to an ammonia-ammonium chloride solution of the metals, using Erichrome Black T as an indicator ( 2 ) . Cheng et al. have been able to determine calcium, magnesium, and iron in limestone with Versenate by varying the buffer conditions (1). The over-all versatility, sensitivity, and general convenience of Versenate as a volumetric reagent are, however, dependent upon the means of end-point detection used for the various titrations. Pribil has described a potentiometric and amperometric procedure for the determination of several cations ( 3 , 4). The use of a spectrophotometric end point for the determination of iron( 111), copper(II), and nickel(I1) with Versenate has been dwcribed*(6).
196
ANALYTICAL CHEMISTRY
In an effort to extend the versatility of ethylenediaminetetraacetic acid (Versenate) as a volumetric reagent, the ultraviolet region of the spectrum has been studied for use in the spectrophotometric determination of the end point of Versenate titrations. A procedure is given for the use of the spectrophotometric end point in the determination of the combined content of calcium, magnesium, cadmium, and zinc or for any one of these metals when present singly in ammonia-ammonium chloride buffered
In an effort to extend the use of the spectrophotometric end point in Versenate titrations, the ultraviolet region of the spectrum has been studied. A procedure is described for the determination of calcium, magnesium, cadmium, and zinc in ammonia-ammonium chloride solutions: calcium in the presence of magnesium; and cadmium in the presence of zinc. In addition, a volumetric method haa been developed for the determination of zirconium, in which an excess of Versenate is added to a zirconium solution and the excess Versenate titrated with standard iron( 111) solution using a spectrophotometric end point of the iron(II1)salicylic acid system. The spectrophotometric end point is not only more versatile than the use of indicators in Versenate titrations but is also more sensitive. Versenate solutions as dilute as 0.001M will give a very sharp end point even when the total volume of the eolution being titrated exceeds 100 ml. ii preliminary investigation of the natore of the titaniumVersenate and titanium-hydrogen peroxide-Versenate complexes was made. The use of either of these complexes for analytical purposes does not offer any decided advantages over existing methods for determining titanium.
solutions; calcium in the presence of magnesium; and cadmium in the presence of zinc in sodium hydroxide solutions. A method for the determination of zirconium involves the titration of an excess of Versenate with iron(II1) solution. A titaniumperoxide-Versenate complex was also found that was stable at low pH levels. The sensitivity of the spectrophotometric end point allows the use of Versenate solutions as dilute as 0.001M or less, even when the total volume of the solution being titrated is 100 ml.
Standard Magnesium, Cadmium, and Zinc, 0.01M. Solutions of these metals were prepared by dissolving accurately weighed quantities of the pure metals in hydrochloric acid and diluting the first solution to 1 liter. Standard Iron(III), 0.018M. Reagent ferric ammonium sulfate was dissolved in water to which a few drops of sulfuric acid was added. The resulting solution was standardized against the Versenate solution. Buffer Solution, pH 10. Sixty grams of ammonium chloride were dissolved in water to which were added 570 ml. of concentrated ammonia, and the solution was diluted to 1 liter with water. SPECTROPHOTOMETRIC DETERMINATION OF END POINT
Spectra of Versenate and of magnesium Versenate, both a t a pH of 10, are given in Figure 1. The spectra of zinc, cadmium, and calcium Versenate are similar to that of the magnesium chelate.
2.0L 1.6
APPARATUS
A Beckman model DU spectrophotometer was used for all the titrations with Versenate. The only modifications required for this instrument was the usual Beckman IO-cm. cell compartment, a quartz titration cell, and a cover for the cell compartment. The T-shaped titration cell, with a 7-cm. light path, was similar in every respect to the cell described previously (6) except that the cell was made from quartz. This quartz cell allowed titrations to be made a t wave lengths as low as 215 mp. A Bakelite cover was made to fit over the 10-cm. cell compartment of the spectrophotometer. In the center of this cover, a 3.5-em. hole was made which allowed the lower portion of the titration cell to rest on the base of the cell compartment while the neck of the cell protruded above the cover. The motor-driven stirrer and 10-ml. microburet were the same as previously described (6). REAGENTS
Standard Versenate, 0.0134M. Approximately 5.0 grams of disodium dihydrogen ethylenediamine tetraacetate dihydrate were dissolved in water and diluted to 1 liter. This solution was then standardized against the standard calcium solution using the spectrophotometric end point. Standard Versenate, 0.0013-lf. Tu-enty-five milliliters of the 0.01338M Versenate were diluted to 250 ml. with redistilled water, and the molarity was checked against standard calcium solution. The molarity found by this procedure was 0.001321, indicating that very dilute Versenate solutions should not be prepared by dilution from an aliquot of standard Versenate without further standardization. Standard Calcium, 0.01M. An accurately weighed amount of reagent grade calcium carbonate (1.0 gram) was dissolved in hydrochloric acid and the resulting solution diluted to 1 liter. Standard Zirconium, 1.035 mg. per ml. Ignited zirconium dioxide (99.9yo purity) was dissolved by a potassium bisulfate fusion. The final solution was 3.6,V in hydrochloric acid. Standard Titanium, 1.043 mg. per ml. Ignited pure titanium dioxide was taken into solution by a potassium bisulfate fusion. The final solution was 3 . 8 5 in hydrochloric acid.
218
222
226
230
234
238
242
246
250
254
WAVE LENGTH, m r
Figure 1. Absorption Spectra of Versenate and Magnesium Versenate A. B.
1.33 X 10-8 Magnesium Versenate, pH 10 1.33 X 10-8 Disodium Versenate, pH 10
The spectrum of the Versenate a t pH 10 is due to the HY --and Y - - - - anions of the Versenate. A similar spectrum is exhibited by Versenate solutions from a pH of about 7 to very alkaline media. .4t pH levels lower than 6.0, Versenate solutions show absorption spectra similar to that of the magnesium Versenate (curve A in Figure 1). It is possible, therefore, to make use of the large absorbancy exhibited by these anions of Versenate to determine end points spectrophotometrically in the titrations of several cations in alkaline solution with Versenate. Since the cations of magnesium, calcium, cadmium, zinc, and their Versenate chelates show little absorbancy a t wave lengths above 222 mp, the end point in the Versenate titration will be indicated by a sudden increase in the absorbancy of the solution caused by an excess of the HY --- or Y ---- Versenate anions. A typical titration curve for the titration of 83y of magnesium with 0.0013M Versenate a t a wave length of 222 mp (magnesium concentration was 0.83 p.p.m.) is given in Figure 2. In titrations with 0.01M Versenate, a wave length of 228 mp was
197
V O L U M E 2 6 , N O . 1, J A N U A R Y 1 9 5 4 employed when the p H 10 buffer was used. Although this is not the absorption maximum of the Versenate, the sensitivity is great enough a t this wave length to give very sharp titration curves. In the Versenate titrations in more alkaline solutions and in alkaline citrate and cyanide solutions, the large absorbance of these reagents would not allow the adjustment of the spectrophotometer to zero absorbancy a t the usual 228 mp, so that wave lengths as high as 236 mp were used when necessary. These wave lengths still gave very satisfactory titration curves. In titrations of cations a t p H levels much lower than 7, it is no longer possible to use the HY --- or Y ---- absorption, so that other conditions must be found. In some cases it is possible to follow the course of the titration a t a wave length where the metal Versenate shows a different absorbancy from the free metal ion and from the excess Versenate. An example of this type of titration curve is the manganese(I1)-Versenate titration. At a wave length of 241 mp and a t a p H of 5.6 the manganese(I1)Versenate shows a stronger absorption than the free manganese(11)ions or the free Versenate. No further study was made of this system although the graph of absorbancy us. milliliters of Versenate was a straight line and could undoubtedly be employed for the titration of manganese(I1) solutions with Versenate. A4further spectrophotometric end-point procedure may be used for cations that show no spectral change when chelated with Versenate. This method would involve the addition of excess Versenate to the cation solution under question and then backtitration of the excess Versenate with another cation that will give a sharp end point while still not displacing the original metal Versenate chelate. This method has been used by Piibil for the determination of aluminum, copper, cadmium, zinc, nickel, and lead by titration of the excess Versenate with iron(II1) and use of a potentiometric end point ( 3 ) . Similar titrations using the spectrophotometric end point of the iron(II1)-salicylic acid system (6) should be possible, and would have the advantage that the exact end point would not have to be determined directly. This is a decided advantage in many titrations of this nature, since there is often a tendency for a slow equilibrium in the region of the end point. The volumetric determination of zirconium by a back-titration of the excess Versenate with standard iron(II1) solutions is possible by this spectrophotometric method. This back-titration is as fast and convenient as a direct titration and gives a very sensitive end point. PROCEDURES
Standardization of Versenate. In the standardization of the Versenate solution with calcium, a 5- or 10-ml. aliquot of the standard calcium solution is ipetted into the titration cell, 2 ml. of pH 10 buffer is added, a n 8 the solution is diluted to 90 to 100 ml. with water. The wave, length used for the titrations with
0.013M Versenate is 228 mp; for more dilute Versenate solutions-i.e., 0.001M-a wave length of 222 mp is employed. With the spectrophotometer set a t the proper wave length, the usual spectrophotometric end point procedure (5, 6) is then used in the titrations of the calcium solution with the Versenate. When titrations are carried out with dilute Versenate solutions a blank should be run on the buffer and distilled water. Even with the 0.013M Versenate, this blank is sometimes of the order of 0.01 ml. or more so that it is advisable to check this blank periodically. Determination of Calcium, Magnesium, Cadmium, and Zinc. The procedure for the titration of these cations is the same as for the standardization of Versenate with calcium. Determination of Calcium in the Presence of Magnesium. Five milliliters of a citric acid solution (0.065 gram of citric acid per ml.) and 10 ml. of 6N sodium hydroxide are added to the titration cell with enough water to make a total volume of about 90 ml. The wave length of the spectrophotometer is set a t 234 mp and the solution titrated with the standard 0.013.W Versenate, using the usual spectrophotometric end-point procedure. The end point in this procedure will give the amount of blank correction for the impurities present in the reagents. Sfter the end point has been passed and recorded, as indicated by the absorbancy of the solution, the calcium and magnesium solution is added to the titration cell and the Versenate titration continued until the second end point is reached. The amount of calcium is equivalent to the difference between these two end points. When several determinations of calcium are going to be made it is not necessary to determine the blank correction for every titration, provided the sodium hydroxide and citric acid are carefully measured so that the blank correction may be measured once and will remain the same. The blank should, in any case, be checked periodically. Determination of Cadmium in the Presence of Zinc. Ten milliliters of 6 N sodium hydroxide, 1.0 ml. of 10% potassium cyanide solution, and 80 ml. of water are added to the titration cell and the blank correction is determined on these reagents as for the calcium in magnesium titrations. The wave length employed in this titration is 236 mp. After this end point has been passed and recorded, the solution containing the cadmium and zinc is added to the titration cell and the Versenate titration continued as before until the second end point is reached. The amount of cadmium present is equivalent to the difference between these two end points (the blank correction end point and the final cadmium-Versenate end point). Determination of Zirconium. Ten milliliters of standard Versenate solution and 15 ml. of a sodium acetate solution (0.05 gram of sodium acetate per ml.) are added to the titration cell followed by an aliquot of the zirconium solution. This aliquot should not contain more zirconium than can be chelated with the amount of Versenate added to the solution-Le., not more than 80 mg. of zirconium per milliequivalent of Versenate added. To the titration cell is then added 1.0 ml. of a 6% salicylic acid solution in acetone, followed by enough water to make the total volume 85 to 90 ml. Then 3N ammonium hydroxide is slowly added to this solution with stirring until the pH is adjusted to 4.0. (The amount of 3N ammonium hydroxide required for this pH adjustment may be determined on a separate aliquot of the zirconium.) The wave length of the spectrophotometer is set a t 520 to 525 mp and the excess Versenate titrated with standard iron(II1) solution using the spectrophotometric end point of the iron(II1)-salicylic acid system (6). The amount of excess Versenate is obtained from the usual plot of absorbancy vs. milliliters of iron(II1) added. The zirconium-Versenate complex corresponds to a 1 to 1 ratio of zirconium to Versenate. DISCUSSION
0 2
0 6
10
14
18
2 2
2 6
30
3 4
38
0.001 3M VERSENATE, MI.
Figure 2.
Titration of Magnesium with Standard Versenate
Wave length, 222
mp;
reagent blank given by first end point
The results in the standardization of the Versenate solution with standard calcium using a spectrophotometric end point are given in Table I. The average deviation from the mean in both of the standardizations of 0.013 and the 0.0013-Lf Versenate is of the order of 0.10%. The results of the separate determinations of magnesium, cadmium, and zinc in an ammonia-ammonium chloride buffered solution of p H 10 with standard Versenate are given in Table 11. The total volume of these metal solutions was between 90 and 100 ml., so that the final solution varied from 0.8 to 100 p.p.m. in the metal ions. The average error for the titration of cadmium, magnesium, and zinc with the 0.013M Versenate wa8 about 2 parts per thousand, while for the titrations with 0.0013M Versenate the average error was slightly higher (4.5 parts per thousand). The
198
ANALYTICAL CHEMISTRY
Table I. Ml.
Standardization of Versenate Solutions with Standard Calcium Solutions Calcium Molarity
M1.
Versenate Molarity
Deviation
from Mean %
4.987 4.987 4.987 10.015 10.015
0.01030
3.840 3.835 3.838 7.728 7.697
0.01338 0.01340 0.01338 0.01335 0.01340
0.03 0.10 0.05 0.25 0.16
4.987 4.987 10.015 10.015 10.015 10.015
0.001008
3,808 3.806 7.637 7.628 7,643 7,646
0.001320 0.001320 0.001321 0.001323 0.001320 0.001320
0.05 0.05 0.03 0.17 0.05 0.05
precision in these titration? wa.q much hetter than indicated bjthe average error figure. Although a study was not made of the interferences from other cations in the determination of cdcium, magnesium, cadmium, and zinc, these interferences could be divided into two classes: those compounds that did not chelate with Versenate but which showed large absorbances a t the wave lengths used, and those compounds that do form stable chelates with Versenate. The first type of interference would be evident as soon as the cation solution was added to the titration cell hy the inability to obtain a
the final solution should be approximately 0.6N in sodium hydroxide. Sodium hydroxide was found to be much superior to potassium hydroxide. When larger amounts of magnesium are present, the results of the calcium titrations are always low. This error is undoubtedly due to the coprecipitation or adsorption of the calcium on the precipitated magnesium hydroxide. The addition of citric acid to the calcium-magnesium solution decreases this error considerably, although even with citrate present the titration of 2 mg. of calcium in the presence of 2 mg. of magnesium was 1.3% low. This error is, however, relatively constant for given amounts of calcium and magnesium, so that a correction could be made for this error. Because of the amounts of sodium hydroxide and citric acid present, a considerable blank is caused by the impurities even when reagent grade chemicals are used. It is therefore important to run a blank on the reagents. The results of the titrations of calcium in the presence of magnesium are given in Table 111.
Table 11. Determination of &.lagnesium,Cadmium, and Zinc with Standard Versenate Cation Magnesium
Molarity of Versenate 0.01338 0.001321
Cadmium
0,01338
0.001321
Zinc
0.01338
0.001321
Sample, Mg. Taken Found
Error,
2.062 2.062 2,062 0.2062 0,2062 0.2062 0.0830
0.11 0.24 0.29 0.05 0.34 0.58 0.91
10.84 10.84 10.84
i0.84
1.089 1.089
6.210 6.210 6.210 6.210 0.6210 0.6210
2.058 2.057 2.056 0.2061 0.2069 0.2074 0.0837
10.80
10.87 10.87 10.87 1.082 1.084 6,210 6.223 6,219 6.228 0.6224 0.6232
%
0.34 0.24 0.25 0.25 0.64 0.46 0.00 0.21 0.14 0.29 0.24 0.33
zero absorbancy reading. If only small amounts of these interferences were present, so that the absorbancy reading was not greatly increased, this type of interference would not cause trouble. Many of the transition elements have strong absorbances a t these wave lengths and would thus interfere when present in large amounts. Interfering metals of the second class would chelate with the excess Versenate, thus preventing the HY--- and Y - - - Versenate anions from indicating the end point. In many cases, however, these interferences may be avoided by the addition of cyanide or other complexing agents to the titration cell. I n strongly alkaline solutions. calcium may be titrated with Versenate without interferences from magnesium. Cheng et al. (1)have made use of this in the determination of calcium in limestone using murexide as indicator. The murexide end point in strongly alkaline solutions is, however, not as distinct as the Eriochrome Black T end point in ammonia solutions. The spectrophotometric procedure, on the other hand, not only gives a very sharp end point for the titration of calcium in the presence of magnesium but may also be used for the determination of cadmium in the presence of zinc when similar alkaline Bolutions are employed. In the determination of calcium in the presence of magnesium,
WAVE LENGTH, mp
Figure 3. A. B.
Spectra of Titanium-Peroxide-Versenate Systems
1.01 X 10-8 M titanium 0.3% hydrogen peroxide and pH 1.7 1.01 X 10- a M titanium: 0.3% hydrogen peroxide: 1.33 X 10M Versenate, and pH 1.7
In strongly alkaline solutions (0.6.V sodium hydroxide) zinc will form the zincate ion which is not chelated with Versenate. Cadmium precipitates in alkaline solutions as the hydroxide which shows no amphoteric properties. If this hydroxide precipitate is prevented by the addition of a small amount of cyanide, it is then possible to titrate alkaline cadmium solutions with Versenate without interference from considerable amounts of zinc. The results of the determination of cadmium in the presence of zinc are given in Table IV. . Although the determination of zirconium with Versenate involves a back-titration of the excess Versenate with iron(II1) solution, it is as fast as a direct titration and is one of the few volumetric methods for zirconium. The ironsalicylic acid end point a t a wave length of 520 to 525 mp gives a very sharp titration curve that requires very few absorbancy readings in order to extrapolate back to the exact end point. The optimum pH for the titrations is 4.0. In the titrations of zirconium that were carried out a t pH levels between 1.0 and 2.0 the results were consistently low, although there seems to be little tendency for the iron( 111) to displace Versenate from the zirconium-Versenate chelate. The results of the zirconium titrations are given in Table V. Titrations similar to the back-titrations of zirconium were tried on titanium(1V) solutions. However, in the back-titration of the excess Versenate with iron( 111) solution, the iron(II1) not only chelated with the excess Versenate, but also displaced the Versenate from the titanium-Versenate chelate. This displacement was rather slow, requiring 10 to 20 minutes for equilibrium to be reached when the solution had a pH of 2, and even longer periode-
199
V O L U M E 2 6 , N O . 1, J A N U A R Y 1 9 5 4 Table 111. Determination of Calcium in Presence of Magnesium Calcium, h l g . Found 2.014 2.013 2.016 2.014 2.014 2,017 1.984 2.014 1.990 2.014 1.986 2.014 1.979 2,014 4.052 4.044 4.011 4.044 3.991 4.044 4.035 4,044 3.10 mg. of zinc added.
Taken
5
Magnesium Added, Mg.
Error
0.00
0.05 0.10 0.15 1.49 1.19 1.39 1.73 0.21 0.82 1.31 0.22
70
0.41 0.83 2.07 2.07 1.07 4.14 0.41 2.07 4.14 0.005
of time a t a p H of 4. At least 99% of the Versenate was displaced from the titanium-Versenate chelate a t a pH of 2. It was possible to titrate zirconium in the presence of smaller amounts of titanium with a fair degree of accuracy if equilibrium was reached in the back-titration with iron(II1) before recording the absorbancy readings. However, the time required for such a titration was over 30 minutes, and the results indicated the necessity of a carefully standardized procedure for satisfactory results. For example, after excess Versenate was added to a solution containing 1.94 mg. of titanium(1V) and 5.164 mg. of zirconium, the pH was adjusted to 4.0 and the excess Versenate n a s backtitrated with standard iron(II1) solution. The amount of zirconium found was 5.197 mg., or an error of 0.64%. On the other hand, if, during the back-titration of the excess Versenate in a similar solution of titanium and zirconium, 1 ml. of excess iron(II1) was added and this solution was allowed to stand 30 minutes before titrating the excess iron, the zirconium found was 6.3% low. There is also a tendency for the titanium to precipitate from solution after the iron has displaced the Versenate 0 from the titanium-Versenate chelate, so that the ahsorbanci- readings are not too stahle
peroxide complex was capable of reacting with Versenate to form another complex ion that contained titanium, peroxide, and Versenate. Furthermore, from the titrations it was apparent that the titanium-hydrogen peroxide-Versenate complex or chelate was more stable than the original titanium-Versenate chelate. It was possible to follow the course of the formation of the titanium-hydrogen peroxide-Versenate complex by a spectrophotometric titration of a titanium-hydrogen peroxide solution with Versenate a t a wave length of 450 mp and a t a pH of 1.9 (hydrogen peroxide concentration was 0.3%). This titration gave a good titration plot and an end point that corresponded t o a complr’x having a one to one titanium to Versenate ratio.
Table V. Titration of Zirconium with Standard Versenate -~ Zirconium, Mg. Error, Taken Found 72 5.164 5.164 5 . I64 5 164
a
5.163 5.144 5.129 5.147
0.02 0.39 0.67 0.33
3.010“ 5.037a
3.0 2.5
p H of final soilltion was 1 .L’
.llthough titanium can be determined when hydrogen peroxide is present by either a direct titration with T’ersenate using a wave length of 450 mp or by a back-titration of excess Versenate with iron(III), there seem to be few advantages of these methods over other exiqting volumetric and photometric procedures for titanium 4CKNOW LEDGIIEVT
The financia! aid of a predoctoritl fellowship to the first author from the Eastman Kodah C’o is gratefully acknowledged LITERATURE CITED
-
_ _
Table IV.
_.
-
~
Determination of Cadmium in Presence of Zinc
Cadmium, RIg. Taken Found
Zinc Added, RIg.
Error.
A
(1)
Cheng, K. L., Iiurtz, T., and Bray, R. H.. ANAL.CHEM.,24, 1640 (1952).
(2) Kinnunen, J., and Merikanto, B., Chrmist Anulyst, 41, 76 (1952). (3) Pribil, R., Koudela, Z., and Matyska, B., Collection Czechosloo. Chein. Communs., 16, 80 (1951). (4) PTibil, R., and hlatyska, B.. Ibid., 16, 139 (1951). ( 5 ) Sweetser, P. R., aiid Bricker, C . E., ANJIL. CHEM.,24, 1107 (1952). (6) Ihid., 25,253 (1953). RECEIVED for review N a y 27, 1933. .Irreptrd Octohi*r 27, 1953
Picrates of Alkylpyridines-Correction I n an effort to reduce the stability of the titanium-Versenate chelate further, hydrogen peroxide was added to the titanium solution. It mas hoped that by forming the titanium peroxidd complex the titanium-Versenate chelate would not form and no new chelate containing Versenate would be produced. Homever, when an excess of Versenate was added to an acid titanium solution containing 0.3% hydrogen peroxide and the p H of this solution mas adjusted to 2 to 4, iron(II1) was able to react only with the excess Versenate. The titanium-Versenate chelate in the presence of hydrogen peroxide was now stable enough to prevent the iron(II1) from displacing the Versenate from this complex. The amount of titanium found by this method, however, was slightly low (0.7 to 1.5Yo), indicating that perhaps a small amount of the Versenate was still displaced by the iron(II1). The spectra of the titanium-peroxide system and the titaniumhydrogen peroxideversenate system (Figure 3) showed that the normal titanium-hydrogen peroxide maximum of 390 mG ( pH 1.7) was shifted to a wave length of 365 mp but the shape of the curve remained the same. This would indicate that the titanium-
Work subsequent to the publication of “Picrates 01 hlkylpyridines. Identification by X-Ray Diffraction Powder Patterns” [Janz, G. J., and Solomon, Raymond, AXAL.CHEM.,25, 454, 1775 (1953)l has shown that the data cannot be used ae standards of reference. The target of the chromium tube suffered contamination with tungsten and iron. Thus, rather than monochromatic Cr-K, radiation, the spectrum contained, in addition, the Cr-Kp, Fe-K,, Fe-Kgl, and W-L,, TV-Lp,, W-Lgz lines. The powder patterns contained many extraneous lines leading to erroneous values for the interplanar spacings, and relative intensities. While these data can readily be corrected for the erroneous d spacings, it is not possible to correct the relative intensities of the lines. Further work, using pure Cr-K,, or Cu-K, radiation is necessary if data for use as absolute reference standards for the alkylpyridine picrates are desired. GEORGEJ. JANZ Department of Chemistry Rensselaer Polytechnic Institute Troy, N Y.
