i i fast enough that the pyridine hemichrome never shows up because it is all converted to cyanopyridine hemichrome. Cyanide Hemichrome in Pyridine. If a solution of hemin (8.1 X 10-4Jf)
and pyridine (2.0,M)also containing the standard amounts of alcohol and electrolyte is made 0.25X in sodium cyanide, only the cyanide hemichrome is found on the cathodic branch of the sweep (Figure 5 ) . On the anodic branch essentially only the cyanopyridine hemochrome is available for oxidation. Evidently the reaction
[‘I
Fe+2 - P
1‘
+ CN-
chrome and cyanopyridine hemichrome (second and subsequent scans) may be noticed. This shows up only a t the hignest scan rate used. The data for the high cyanide solution is recorded in Table V. Since the peak current for both the reduction of cyanide heinichrome and the oxidation cyanide hemochrome can be measured a t 4.66 volts/sec., it is possible to apply the method of Nicholson and Shain (6) to the measurement of k,, for Reaction 5 . It was found that kf, = 20 see.-’. Obviously since only one ratio of anodic to cathodic peak current was measured, this calculation is not accurate. It should be viewed rather as an estimation of kf, and is probably within a factor of two of the real value.
Table V. Cyclic Voltammetric Data for Cyanide Hemichrome in Pyridine
Scan rate, volts/sec.
Peak currents, pa. (iAQ (i.)z 0.0258 2.00 ... 2.00 0.246 6.20 ... 6.20 0.485 8.90 . . 8.90 4.66 32.00 16.0 16 0 Subscript 1 refers to the peak for the reduction of cyanide hemochrome and 2, to the peak for cyanopyridine hemochrome. ic
5
(6) Nicholson, R. S., Shain, I., ANAL. CHEM.36, 706 (1964). (7) Ibid., 37, 178 (1965).
LITERATURE CITED
(1) Clark, W. H., “Oxidation Reduction
(5)
is quite fast but in this solution the results of Reaction 3 are not evident. If the scan rate is great (Figure 6) the transient existence of cyanohemo-
Potentials of Organic Systems,” Chap. 8, V7illiams and Wilkens, Baltimore, Md., 1960. (2) Davis, D. G., Martin, R. F., unpublished experiments. (3) DeFord, D. D., Division of Analytical Chemistry, 133rd Neeting, ACS, San Francisco, 1958. (4) Falk, J., E.:! “Porphyrins and Metalloporphyrins, Elsevier, ilmsterdam, 1964. ( 5 ) Jordan, Joseph, Bednarski, T. X., J . Am. Chem. SOC.86, 5690 (1964).
13, 349 (i948j. (11) Shank, J., Clark, W. H., J . Biol. Chem. 171, 143 (1947).
(12) Urry, D. W., Eying, H. Proc. Satl. A c a d . Sei. C.S. 419, 253 (1963).
RECEIVEDfor review August 9, 1965. Accepted Sovember 24, 1965. Diviiion of ilnalytical Chemistry, 149th lleeting, ACS, Detroit, Mich., April 1965. Investigation supported by Public Health Service Grant .43108248 from the Satioiial Institutes of Arthritis and Metabolic Diseases.
Polarographic Reduction of Oxygen in Dimethylsulfoxide EDWARD 1. JOHNSON, KARL H. POOL, and RANDALL E. HAMM Department o f Chemistry, Washington State University, Pullman, Wash.
99 7 6 3
-0.65 volt and attaining a limiting b The reduction of oxygen in the current a t -1.20 volts (us. S.C.E.). In aprotic solvent dimethylsulfoxide dilute solutions two waves were ob(DMSO) has been shown to proceed tained with Ell2 values of -0.72 and through two steps at the dropping- 1.20 volts, respectively. The second mercury electrode. The first step is 2 e = 02-. wave was approximately 75y0 of the due to the reaction 0 height of the first wave. When 0.1M The second wave is not as high as HzSOa or a n acetate-acetic acid buffer the first, but has been represented by was added, only one wave was observed e cation = the reaction 0 2 with El,nof -0.4 volt. peroxide product. This second wave Recently llaricle and Hodgson (4) has been shown to be markedly and Peover and White ( 5 ) , on the basis shifted by change in the cation of cyclic voltametry at a platinum present in the supporting electrolyte. electrode, have shown that the first The height of the second wave changes reduction wave of oxygen in DMSO with change of cation in supporting corresponds to a one-electron process. electrolyte. The diffusion coefficient This observation has been checked by of oxygen in DMSO has been calcuthe use of a x . polarography by Peover lated from the data obtained.
+
+ +
K
(3) observed that in a 0 , l X NaC104 solution of DMSO saturated with oxygen gas, only one wave was obtained starting at OLTHOFF AND REDDY
and White (5) and by the application of epr by Llaricle and Hodgson (4) and by Slough ( 7 ) . When Maricle and Hodgson added phenol as a proton source, the reduction wave was shifted anodically and was approximately doubled in
height. They suggested that the presence of phenol converted the O2 reduction into a two-electron reduction. They also reported that the second reduction wave of O2 was not reproducible and suggested that it varied with the concentration of acidic impurities in the DMSO. The work reported in this paper is a n investigation of the acidic impurity effect and further investigation of the reduction in the presence of proton sources. EXPERIMENTAL
Materials. Purified DMSO was obtained from the Matheson Co. Although the supplier reported t h a t the water content was not greater than 0.057,, samples of this solvent were held over sodium hydroxide for 3 hours at 90’ C. and were then vacuum distilled with only the center cut taken for use. The (C2H5)JC104 (TE-iP) was prepared from (C2HJ4NBr VOL. 38, NO. 2, FEBRUARY 1966
183
obtained from the Matheson Co., according to the method of Kolthoff and Coetzee (1). Sodium perchlorate obtained from G. Frederick Smith Chemical Co. and potassium nitrate obtained from Mallinckrodt were dried at 110' C. prior to use. LiC104 was prepared by neutralizing Li2C03 with HC104 and drying the crystalline product a t 160' C. The KOz of 99% purity was used directly as obtained from K and K Laboratories. Oxygen and nitrogen gases, obtained from Industrial Air Products Co. were passed over anhydrous CaCl? and through concentrated bubblers to dry them before use. To prepare the DMSO solutions of magnesium, calcium, and strontium salts, the respective nitrate salts obtained from llallinckrodt were dissolved in DXSO and placed on a steam bath for 2 days. The crystalline salts were then placed in a vacuum desiccator for 3 additional days a t 50' C. The prepared salts were kept in a vacuum desiccator. Stock solutions of these salts were prepared in DMSO and standardized bv comulexometric EDTA titrations. The HC1-DMSO solution was prei,ared bv bubbling dried HC1 pas. bbtained"from the Xlatheson Co., k t o an ice-cooled solution of DMSO. To standardize this solution, 10-ml. aliquots of the solution were diluted to 50 nil. with water and were titrated with standard sodium hydroxide solution. The normality was 0.894. Procedures. To run polarograms of DMSO saturated with oxygen, a DMSO solution was made 0.1-11 in T E l P by using material that had been dried in a vacuum oven a t 60" C. for 1 week. The solution was then saturated with oxygen by passing the dried gas through the solution in a glove box. The results of these runs compared exactly with those in which DMSO obtained from the supplier was used without further purification.
Table
I.
Wave Heights for Reduction of Oxygen in DMSO
First and second wave in presence of varying oxygen concentration Concn. of 0 2 , First Second mmole/liter wave, pa. wave, pa. Supporting electrolyte 0.10M TEAP 1.65 1.08
8.7 5.7
2.4 2.8
Supporting elect'rolyte0.10M NaC104 1.71 1.08 0.47 0.17 0.11
184
9.0 5.7 2.5 0.9 0.6
ANALYTICAL CHEMISTRY
6.8 5.1 2.4 0.9 0.6
-e d
Y
5- - 1.7 i I
- 1.3 PO
40 60 Concn., m M
80
100
Figure 1. Half-wave potential for the second wave of reduction of dissolved oxygen in dimethylsulfoxide which was 0.1 M with tetraethylammonium perchlorate where NaC104 or KNO, was substituted for a small part of the
(CdshNC104
the second wave was usually somewhat less than the first wave but depended upon the concentration of oxygen in the solution. h deaerated DMSO solution that was made approximately 1.5 mM in KO2 yielded a polarogram with an anodic wave having a half-wave potential of - 1.15 volts. The current reached zero following this wave and remained there until a second wave appeared a t Eli2 = -2.50 volts. This agrees with the assignment of the first wave as a oneelectron reduction to superoxide anion made by Maricle and Hodgson (4). When a HCI-DMSO solution was added to the solution of oxygen in DMSO, the first wave shifted anodically to give a EL/?of -0.84 volt. The second wave started to disappear and the first wave increased in height with the addition of HC1 in DMSO. Upon adding about 4 equivalents of acid per mole of dissolved oxygen, only the first wave was observed a t about twice its original height. Further addition of acid no longer changed the height of the first wave. When varying amounts of 0.1M solutions of XaC104 or KNO, in DMSO were added to the DMSO solution which was 0.1M in TEAP and the resulting mixture was saturated with dry oxygen, the first wave remained unchanged and the second wave shifted anodically and its location depended upon the concentration of the metal ion present as illustrated in Figure 1. The height of the second wave was a function of the particular cation in the supporting electrolyte, but the height of the second wave relative to that of the first wave became larger as the amount of oxygen dissolved was decreased. This effect is shown by the data in Table I. The effects of adding Li+, Mg+2,Ca+*, and Srf2 were different because the second wave disappeared when the
Apparatus. A Leeds and Northrup Electrochemograph Type E was used t o record all current-potential curves. Standard H-cells supplied by the Sargent Co. were used. One branch of the cell was used as the dropping mercury electrode (D.M.E.) compartment, the cross arm contained a glasswool plug and the other branch of the cell contained a dry DMSO solution saturated with KC1 and an AgC1-Ag electrode and was sealed off from the air. The potential of the AgC1-hg electrode in DMSO was completely reproducible and when measured against a Beckman fiber-type saturated calomel electrode (S.C.E.) with a vacuum tube voltmeter was $0.30 volt. All cell potentials reported in this paper are potentials measured against the saturated KC1(DMSO)/AgCl-Ag electrode and should be made 0.30 volt less negative to be referred to S.C.E. Measurement of cell resistances using varying supporting electrolytes showed that errors due to IR drop were less than 0.01 volt in all cases. The D.M.E. used had an m value of 1.03 mg./second and an open circuit time of 5.10 seconds. The half-wave potentials of O2 in a 0.1M solution of TEAP in DMSO taken in Y the described cell were -1.15 and b, - 1.9 -2.50 volts. The concentration of O2 was determined using the Winkler method (8) as modified by Pomeroy and Kirschman (6). Another modification found necessary was, after the addition of H2SOa to the sample, transferral to a. 125-ml. Erlenmeyer flask containing 40 ml. of ice-cooled water to prevent loss of - 0.8 iodine vapor due to the heat of dilution. PO 40 60 80 100 Concn., m M Freshly prepared starch indicator was added and the solution was titrate: Figure 2. Half-wave potential for with standard thiosulfate. the first wave of reduction of disRESULTS
Each polarogram that was run on dissolved oxygen in 0 . M TEAP in DMSO showed two waves with Ell2 values of -1.15 and -2.50 volts. The height of
solved oxygen in dimethylsulfoxide which was 0.1 M with tetraethylammonium perchlorate where lithium, calcium, magnesium, or strontium salts were substituted for small amounts of the ( C ~ H S ) ~ N C I O ~
ing from 0.44 mmole/liter to 1.67 mmoles/liter. The average value of DO,ll2 was found to be 5.31 X with a standard deviation of 0.51 X
10-3. Each of these 10 solutions was titrated amperometrically with HC1 in much the same manner as was used by Kolthoff and Niller (2). A plot of such an amperometric titration is shown in Figure 3. If C represents the bulk concentration, then the condition for the endpoint should be
.t B 2
4 6 Moles H+/Moler Oa
8
Figure 3. Amperometric titration of a solution containing 1.03 mmoles of oxygen per liter of 0.1M tetraethylammonium perchlorate in DMSO with 0.894M solution of HCI in DMSO at a potential of -0.75 volt vs. S.C.E.
From the endpoint of each of the 10 amperometric titrations, the diffusion coefficient of the hydrogen ion was calculated. The average value obtained for DH1I2was 2.07 X 10-3 with a standard deviation of 0.16 X lo+. This gives a value for D H of 4.29 X 10-6 compared to a value of 4.4 x as reported by Kolthoff and Reddy (3). DISCUSSION
metal ion was added and the first wave showed an anodic shift and also increased in height. The shift of the first wave as a function of added Li+, i\Ig+?, Ca+2, and Sr+? is shown in Figure 2. In no case did the first wave become twice as high as the first wave had been, but with the addition of the metal ion the height of the second wave seemed to limit depending upon the ion added. With Li+ the wave became 1.2 times as high; with Ca+2, the wave became 1.5 times as high; and with hIg+2and Sr+2, the wave became 1.6 and 1.4 times as high, respectively. A controlled potential electrolysis a t -1.50 volts in a DMSO solution which was continuously saturated with oxygen and contained an excess of dry HC1 gave H2O2as a product as evidenced by the formation of a bright yellow color when Ti+4was added to the solution. The use of the Ilkovic equation to determine n requires a knowledge of the diffusion coefficient of oxygen in DMSO and this coefficient was not obtainable. Since it has been concluded that the number of electrons involved in the first reduction wave was one, the diffusion coefficient of oxygen was calculated from the experimental data on reduction of 10 different solutions with oxygen concentrations rang-
The results of the KOnexperiment and the production of HzOzin the presence of excess strong acid are in agreement with the reports of other workers who have proposed that the first reduction wave is a quasireversible one-electron reduction. The reduction processes for two waves may therefore be written as: oxygen in dry DAIS0 first wave oxygen second wave
+ HC1
The effect upon the second wave of varying the supporting cation can be explained by the specific action of the metal ion a t the mercury surface to stabilize the peroxide products. The fact that sodium shifts the wave more anodically than potassium (- 1.44 us, - 1.85 volts) may be explained by the fact that the peroxide of sodium is more stable. The variation of the heights of the waves shown in Table I indicates that solubilities of the intermediate product may be involved. The results when lithium, magnesium, calcium, and strontium are added indicate that there is a possibility of getting a stable superperoxide product and that this reaction may be written
02
+ Li+ + 2e
-t
Li02
or
the increased height of the wave may be due to the decomposition reaction
because observation of the change of current as a function of mercury column height indicates that the process a t this point does not have pure diffusion control. Further investigation of this for a number of metal ions is in progress in this laboratory. The difference between the half-wave potential for the second wave as observed by Xaricle and Hodgson (4) in tetrabutylammonium perchlorate and that observed in this investigation in tetraethylammonium perchlorate can be attributed to the specific cation effect noted in this investigation. *is seen in Figure I, only small trace.. of sodium ion would cause change. in the half-wave potential of the second wave. It is probably the change. in content of certain trace metal ion- in the DMSO from experiment to experiment and not the acidic impurities which cause the variability of the half-wave potential of the second wave.
O2+ e s 0 2 O2+ 2H+ + 2e 02-+ 2>1+
+e
-
Ha04
+
11202
LITERATURE CITED
(1) Kolthoff, I . AI., Coetzee, J. F., J . Am. Chem. Soc. 79, 870 (1957). (2) Kolthoff, I. M.,Miller, C. S., Ibid., 63, 1013 (1941). (3) Kolthoff, I. AI., Reddy, T. B., J. Electrochem. Soc. 108, 980 (1960). (4) Maricle, R. L., Hodgson, R. G., ASAL.CHEM.37, 1562 (1965). (5) Peover, AI. W., White, B. S., Chem. Commun. 1965, 183. (6) Pomerov. R.. Kirschman. H. D.. IND. E N G . CH&. AiXAL. ED.17,' 715 (1945). (7) Slough, W., Chem. Commun. 1965, 184 (8) Winkler, L. W., Ber. 21, 2843 (1888). ~
RECEIVEDfor review June 28, 1965. Accepted November 22, 1965. Supported by research funds of Washington State University.
VOL 38,
NO. 2,
FEBRUARY 1966
185
