previous measurements (2) of its chemical shift, and obtain the same results within 0.6 OK. METHOD
A spectrometer with an internal lock was used to obtain the present results. With the spectrometer locked to the CHI protons, the recorder is centered on the OH resonance, and the frequency differenceAVis read directly on a counter. A trace (0.03 by volume) of concentrated aqueous hydrochloric acid was added to the air-saturated methanol or glycol samples (Fisher reagent grade). This caused a complete collapse of the multiplet structure, and sharp lines were obtained at all temperatures. Most likely, even smaller amounts of HCl would be adequate. The addition of HCl does not affect the chemical shift at any temperature, and the sharp lines are a great convenience. The esterification reaction of the HC1 with the methanol is not noticeable over a period of half a year, and the lines remain sharp. Oxygen or nitrogen gas was bubbled through some methanol samples. Within the accuracy of 0.5 Hz, the chemical shift is not affected, at least not at 300 OK. Also, the chemical shift of the air-free methanol and glycol standards supplied by Varian Associates is identical to our air-saturated samples at all temperatures. Moisture has a small effect. Temperatures were measured with a static thermistor sensor (2) in a spinning sample tube. Its resistance was measured with a bridge (Electroscientific Industries 250 DE, Portland, Ore.). Using an external bridge resistor of 100 Q,resistances up to 100 MQ could be measured to better than 0.5 %. At the
melting point of methanol (175.4 OK), the resistance was 22.4
MO. The thermistor was calibrated by comparison with a vibrating quartz thermometer (Dymec 2801 A, Hewlett-Packard), and the calibration agreed with the melting point of methanol (observed 175.3 OK, lit. 175.4 OK) and with the sublimation point of carbon dioxide(obsd 194.5 h 0.2 OK,lit. 194.7 OK). The temperature at the coil is affected only slightly by spinning, and then only at low temperature. When the spinning was stopped at 230 OK, the temperature fell only 0.4 OK. At 390 OK, no change could be observed (reproducibility 0.2 O K ) . The bottom of the sample tube was 10 mm below the receiver coil. In our experience, temperature measurement with the thermistor thermometer is faster and more accurate and reliable than measurement of the methanol shift. For most purposes, it is not necessary to spin the tube. The thermistor thermometer should be inserted to the right depth, since a temperature gradient occurs (0.5 OK/cm at 233 OK). It should be remembered that the probe constitutes a heat leak. At 233 OK, the probe raises the temperature by 1.0 OK. ANTHONY L. VANGEET Department of Chemistry Oakland University Rochester, Mich. 48063 RECEIVED for review January 15, 1970. Accepted February 9, 1970.
Simultaneous Spectrophotometric Determination of Hydrogen Peroxide and Peroxyacids of Sulfur SIR: A few years ago ( I , 2) a cerimetric macro scale method was described for the simultaneous determination of hydrogen peroxide, peroxymono- and peroxydisulfuric acids. The method was based on the selective removal of peroxymonosulfate by arsenious acid. There were some attempts, too, ( 3 , 4 ) to use this procedure for the estimation of minute quantities. We have to say, however, that we were not satisfied with the accuracy offered by our method in micro scale; therefore we tried to find a more appropriate way of analysis. In the following we describe a micro method successfully applied for studying the X-ray radiolysis of sulfuric acid solutions as well as that of sulfuric acid glasses (5). The estimation of hydrogen peroxide can be carried out directly by measuring the absorbance of the peroxo-titanium(lV) complex at 410 mp, which is not interfered by any substances present. Unfortunately, the absorbance of the peroxo complex is not high enough (the molar absorptivity is 718 1. mole-km-l in 0.5M sulfuric acid); consequently, only solutions having [H202]2 lO-5M can be determined with sufficient accuracy. (1) L. J. Cdnyi and F. Solymosi, Acra Chim. Hung. Acad. Sei., 13,257 (1958). (2) L. J. Cshyi, J. Bityai, and F. Solymosi, 2. Analyr. Chem., 195,9 (1963). (3) M. Daniels and J. Weiss, J. Chem. SOC.,1958,2467. (4) I. W. Boyle, Radiarion Res., 17, 427 (1962). (5) L. J. Cshnyi, Final report for G. E. C. grant, Leeds, 1963. 680
*
ANALYTICAL CHEMISTRY, VOL. 42, NO. 6, MAY 1970
The next step of the analysis is the estimation of the sum of hydrogen peroxide and peroxymonosulfuric acid. To this end the sample is added to a mixture of arsenious acid and osmic acid catalyst and after the quantitative reduction of the peroxo titanium(1V) complex and peroxymonosulfuric acid (about 10 min) the excess of arsenic(I11) is determined spectrophotometrically by cerium(1V) sulfate reagent at 320 mp. Finally, the total oxidizing capacity is estimated by using ferrous sulfate reagent, measuring the absorbance of ferric ions formed at 304 mp. EXPERIMENTAL
Apparatus. The absorbance of solutions was measured by Unicam SP 500 spectrophotometer supplied with a thermostated cell housing. The temperature was kept at 25 f 0.15 "C. Reagents. Ordinary distilled water was purified by two successive redistillations. The first redistillation was from an alkaline permanganate medium, the second from a slightly acidic peroxydisulfate medium (0.05M in sulfuric acid containing 2 g of potassium peroxydisulfate per liter). Peroxydisulfate stock solution was always freshly prepared by dissolving K&Os N-free Merck reagent of G.R. grade in triple-distilled water. Peroxymonosulfuric acid solution was prepared from KHSOs c.p. reagent supplied by Laporte Chemicals Ltd., by dissolving it in 0.01M sulfuric acid. Hydrogen peroxide solution was prepared from Merck Per-
Table I. Test Analysis of Standard Solutions (Single Measurements) Difference, 7.8 15.6 15.6 15.6 15.6 15.6 16.2 24.3 24.3 24.3 24.3 24.3 31.2 32.4
17.0 17.0 17.0 17.0 34.0 8.5 25.7 25.7 25.7 25.7 34.2 17.1 17.0 25.7
19.3 19.3 38.6 9.7 19.3 19.3 27.8 27.8 37.1 18.6 27.8 27.8 19.3 27.8
7.6 15.9 15.9 15.7 15.9 15.9 16.7 25.1 24.2 25.1 25.1 24.1 31.1 33.0
17.8 16.7 17.1 16.9 33.6 8.4 25.1 24.9 25.0 25.5 34.7 16.5 16.9 25.6
19.6 19.9 38.7 10.0 20.3 19.8 27.0 29.0 38.2 17.4 25.4 27.8 20.2 27.0
HzOz
HzSOs
HzSzOs
-2.6 +1.9 +1.9 +0.7 +1.9 +1.9 +3.1 +3.3 -0.4 $3.3 +3.3 -0.8 -0.3 +1.8
+4.7 -1.8 +0.6 -0.6 -1.2 -1.1 -2.3 -3.0 -2.7 -0.8 +l.5 -3.5 -0.6 -0.4
-1.55 +3.1
+O. 3 +3.1 +5.2 +2.6 -2.9 $4.3 +3.0 -5.5 -8.7 0.0
+4.7 -2.9
Table 11. Discrepancies Observed Using Mariano's Method (6)
12.3 12.3 24.5 24.5 37.0 37.0 37.0 49.1 49.1 49.1 84.9 84.9 98.0
48.5 48.5 29.1 29.1 29.1 19.3 9.7 77.5 58.1 19.3 77.5 9.7 9.7
27.8 27.8 69.6 69.6 69.6 13.8 27.8 27.8 69.6 13.8 27.8 27.8 27.8
12.0 12.1 23.5 23.1 35.4 34.8 34.7 46.3 47.2 47.6 81.2 80.9 93.7
50.0 50.3 30.7 31.2 31.2 20.1 10.9 80.8 61.7 20.6 82.3 10.8 11.7
hydro1 G.R. diluted by 0.01M sulfuric acid. Other chemicals were also Merck preparates of G.R. grade and used without further purifications. The analyses were carried out in admixtures of peroxides prepared immediately before use. The compositions of the stock solutions of each of the peroxides were checked spectrophotometrically by using ferrous sulfate reagent. All glasswares were cleaned by permanganic acid (finely powdered potassium permanganate dissolved in concentrated sulfuric acid) which was followed by thoroughly rinsing with nitric acid containing hydrogen peroxide, then with tapwater and lastly with triple-distilled water. Procedure. DETERMINATION OF HYDROGEN PEROXIDE. An appropriate size of sample (10-150 p M total oxidizing capacity) is added to a 100-ml volumetric flask containing 10 ml of 10-2M Ti(SO& solution and sulfuric acid sufficient to adjust the acid concentration to OSM. The flask is then filled to the mark. The absorbance of this solution (solution A) is checked at 410 mp, in cells of appropriate path length (2-10 cm) depending on the concentration of hydrogen peroxide, by using a blank containing no peroxy compounds. DETERMINATION OF THE SUMOF THE HYDROGEN PEROXIDE AND PEROXYMONOSULFURIC ACID. Ten ml of 4.5 X 10-4M arsenious acid (in 20 wjw Z sulfuric acid) and 1 ml of 4 X lO-'M osmic acid (also in 20 w/w Z sulfuric acid) and 10 ml of solution A are mixed in a stoppered flask and after 10 min 10 ml of 6 X 10-4M Ce(S04)2in 1M sulfuric acid is added. The difference in the absorbance is measured at 320 mp by using a blank having the same composition but without solution A. A 1-2 cm path length is convenient. Calibration curve offers more reliable data than the use of an apparent molar absorptivity (which amounts to 11800 1. mole-'cm-' relating to peroxy compounds). The blank solution can be stored for a couple of days.
26.1 26.3 66.8 66.7 67.0 13.6 26.0 27.0 65.9 13.6 26.4 27.4 26.6
-2.4 -1.6 -4.1 -5.7 -4.3 -5.9 -6.2 -5.7 -3.9 -3.0 -4.4 -4.7 -4.4
+3.1 -1-3.7 +5.5 $7.2 +7.2 $4.2 +12.4 +4.3 +6.2 t6.8 +6.2 +11.4 f20.6
-6.1 -5.4 -4.0 -4.2 -3.7 -1.5 -6.5 -2.9 -5.3 -1.5 -5.0 -1.4 -4.3
DETERMINATION OF THE TOTALOXIDIZING CAPACITY.I n a stoppered flask 10 ml of 10-2M FeS04 in 20 w/w Z sulfuric acid and 10 ml of solution A are mixed up and after 20 min the absorbance is measured at 304 mp using a 1-5 cm path length. The blank for this measurement should be composed similarly. The apparent molar absorptivity is about 4300 1. mole-1cm-1, but it is better to use a calibration curve. RESULTS AND DISCUSSION
To illustrate the usefulness of this method, some data of single measurements are compiled in Table I. This method furnishes figures for each component better than i3 Z , Very recently a spectrophotometric method was described by Mariano (6) for the analysis of the above system. Her method was based also on the selective reduction of peroxymonosulfuric acid (the recommended waiting time is 10 min) followed by the spectrophotometric determination of hydrogen peroxide with ceric sulfate reagent. Then the dissolved oxygen is removed by bubbling nitrogen through the solution for about 10 min and the peroxydisulfate content as well as the excess of ceric sulfate is estimated spectrophotometrically by using ferrous sulfate reagent. Finally in a new sample the total oxidizing capacity is determined with the aid of ferrous sulfate, by measuring again the absorbance of iron(II1) formed. This fairly simple method attracted our attention and we tried to use it to solve some kinetic problems. Unfortunately, however, we could not obtain sufficiently accurate results. As can be seen in Table 11, regular negative H202- and H2S208(6) M. H.Mariano, ANAL.CHEM., 40, 1662(1968).
ANALYTICAL CHEMISTRY, VOL. 42, NO. 6, MAY 1970
0
681
errors and also regular positive H2S06-errors were found. To explain this experience we assumed that the errors arise from several sources : The rate of reaction between hydrogen peroxide and arsenious acid is rather low. According to Woods, Kolthoff, and Meehan (7) the second order rate coefficient is 1.O X 1. mole-lsec-l at 25 "C. In good agreement with this figure, an 8x conversion of hydrogen peroxide was observed at experimental conditions prescribed by Mariano (10-4M hydrogen peroxide, 1.6 X 10-2M arsenious acid, and 0.4M sulfuric acid). Previously it was pointed out (8) that arsenious acid reacts with peroxymonosulfuric acid fairly quickly. But the reaction is not quantitative within 5 min even at higher concentrations (0.1N) when the acidity is low. Consequently, in a very dilute solution the removal of peroxyrnonosulfuric acid will not be quantitative during a 10-min waiting time. On the other hand it is known (9-12) that at similar conditions a fast reaction takes place between cerium(1V) ions and peroxymonosulfuric acid resulting in the nearly quantitative disappearance of Caro's acid, while a part of cerium(1V) is also reduced. Through the disappearance of cerium(IV), this reaction may result in the reduction of the negative H202error. The reaction between arsenious acid and cerium(1V) is very slow in the absence of catalysts. In the presence of active impurities, however, this reaction could also become the source of positive HzOz-error. Concerning the negative H&08-error, it should be mentioned that according to Woods, Kolthoff, and Meehan (13) (7) R. Woods, I. M. Kolthoff, and E. I. Meehan, J. Amer. Chem. Soc., 86, 1698 (1964). (8) . , L. J. CsBnyi and F. Solymosi, Acta Chim. Hung. Acad. Sci., 17, 69 (1958). (9) L. J. CsBnyi and F. Solymosi, Acta Chim. Hung. Acad. Sci., 13. 19 (1957). (10) Ibid., 15, 501 (1956). (11) L. J. CsBnvi. F. Solymosi, and F. Sziics, Naturwiss., 46, 353 (1959). (12) L. J. Csdnyi and L. Domonokos, Acta Chim. Hung. Acad. Sci., 34, 383 (1962). (13) R. Woods, I. M. Kolthoff, and E. I. Meehan, J. Amer. Chem. SOC.,85, 2385 (1963). ~~
I
Sir: I was exceedingly surprised at the data plotted in Table I1 of Dr. Csanyi's paper and the interpretation which followed. In connection with his remarks on my work ( I ) , I would like to call your attention to the following points. 1. From the data given in such table it is inferred that, together with the [Hz02]and the [H2S208]values, those corresponding to the total oxidizing power of the solutions present, (1) M. H. Mariano, ANAL.CHEM., 40, 1662 (1968).
the reduction of peroxydisulfate by arsenious acid is induced by the reaction between iron(1I) and peroxydisulfate. In the presence of iron(II1) and copper(I1) the eq. of As(II1) oxidized can approach ineq. of Fe(II1) oxidized) finity, i.e, peroxydisulfate will be reduced by arsenic(II1) instead of iron(I1). Since the induced reduction of peroxydisulfate depends on the concentration ratio of [As(III)]/ [Fe(II)] and further, on the concentration of ferric ions, the magnitude of HzSz08-errorwill also depend on the concentration of hydrogen peroxide present as well as on the ferric content of the ferrous sulfate reagent. At low hydrogen peroxide concentrations more cerium(1V) ions remain in the solution after the oxidation; therefore more ferric ion will be formed in the reaction of ceric ions with ferrous sulfate. Consequently, in such cases a greater H2S208aTOr occurs. As according to Mariano's method, the peroxymonosulfate concentration is obtained by difference between the total oxidizing capacity and the sum of [H202] [HzSd&], it is understandable that the peroxymonosulfuric acid concentration can be obtained only with positive error. Disregarding the discrepancies observed, that may arise partly from the slight difference of reagents used, we are of the opinion that the first method is preferable, especially at kinetic runs, because the titanium sulfate is not only a reagent for the determination of hydrogen peroxide, but also a means for quenching the reaction. At the latter procedure, however, such a facility is not offered.
)
+
ACKNOWLEDGMENT
My gratitude is due to Mrs. M. Palotai for her technical assistance and to Laporte Chemicals Ltd., Luton, England, for providing the KHSOs preparation.
L. J. CSANYI Institute of Inorganic and Analytical Chemistry A. Jdzsef University Szeged, Hungary RECEIVED for review May 27, 1969. Accepted November 10, 1969.
also, regular negative errors (there is only one honorable exception, that of the second run where, apparently, an exact determination was performed). If we compare the well behaved T.O.P. values of Table I with those of Table I1 (obtained by exactly the same procedure), we arrive at the obvious conclusion that, in the latter case, between the moment of the preparation of the stock solutions and their measurement, some of the reactants have undergone a decomposition. Now, if the HzOzstock solution was not kept in a sufficiently
Table I. Values Obtained in Simultaneous Spectrophotometric Determinations of H202,H2S208,and H2S05
682
43.1
105.9
82.3
41.7
101.3
90.1
10.5 51.4 94.8
14.4 55.5 75.0
81.7 46.0 20.5
43.9 42.7 43.2 43.6 10.9 51.2 92.0
ANALYTICAL CHEMISTRY, VOL. 42, NO. 6, MAY 1970
106.1 104.2 98.1 101.6 13.9 56.2 76.3
82.3 84.7 91.2 86.6 82.1 44.5 21.5
+1.9 -0.9 +3.5 +4.5 +3.8 -0.4 -2.9
+0.2 -1.6 -3.2 $0.3 -3.5 +1.3 +1.7
0.0 +2.9 11.2 -3.9 +0.4 -3.3 +4.8
