15.0
10.0
5.0
! 0.0
-5.0
-10.0
-0 57
-0.53
-0 51
-0.S3 E
'j.
-3.65
-0.57
-5.
j
C.E., w l t 8
Figure 4. Comparison Of theory (circles) for reduction of cadmium
and experiment
anodically. This effect is reasonable because an increase of y corresponds to a decrease of D E ,in which case the surface concentration of R relative to 0 would increase ( R diffuses into the electrodes more slowly). This causes a Nernstian shift of the waves along the potential axis to more positive values. Since both cathodic and anodic peaks shift simultaneously. _ . the difference of the peak potentials is less affected. Similar interpretations can be given for the influence of I$ and Ex on peak potentials. EXPERIMENTAL RESULTS The calculations described above were evaluated experimentally for reduction of cadmium in aqueous solution and
reduction of sodium in acetonitrile. In both cases excellent agreement was obtained between theory and experiment. The cadmium system, however, is much better characterized in terms of physical constants, and therefore only those results are presented. Experiments were performed using standard equipment and techniques that have been described previously (8). The cadmium system is ideal for checking the calculations (and, implicitly, Reinmuth's approximation) because Stevens and Shain (I) have carefully measured all of the necessary physical constants. Hence, our experiments were performed under precisely the same conditions as those of Stevens and Shain, so that their value of 0.678 for y could be used. A typical experimental curve is illustrated in Figure 4 (circles). This curve was recorded at a scan rate of 21.5 mV/sec with a HMDE of radius 0.0676 cm. The calculated value of C$ in this case is 0.031. To compare the experimental data of Figure - 4 with the new theory, the current function x(r,C$) was calculated from the integral equation for y equal 0.678 and 4 equal 0.031. These values of x(y,#,) were then used with Equation 1 and the other known physical constants to determine the theoretical current in microamperes. This result is represented by the solid line in Figure 4, where it is seen that the agreement between theory and experiment is essentially perfect. Since no normalization was used in this comparison, the agreement is as good as any ever published for cyclic voltammetry, and persuasive evidence of the validity of using Reinmuth's approximation. RECEIVED for review March 7, 1972. Accepted April 26, 1972. Financial support of this research by the National Science Foundation is gratefully acknowledged. (8) R. S. Nicholson, ANAL.CHEM ., 37, 1351 (1965).
Electrochemistry of Lucigenin Kenneth D. Leggl and Donald W. Shive2 Department of Chemistry and Laboratory for Nuclear Science, Massachusetts Institute of Technology, Cambridge, Mass. 02139
David M. Hercules* Department of Chemistry, Uniljersity of Georgia, Athens, Ga. 30601 The electrochemistry of lucigenin has been studied in nonaqueous solvents. The initial reduction of lucigenin has been shown to proceed by a one-electron charge-transfer step, leading to the monocation radical. The charge-transfer step is followed by a rapid disproportion at ion of the radical , yielding d imethylbiacridine as the product. The rate constant for the disproportionation has been determined to be 4.41 x 104M1-1-SeC-'. Oxidation of dimethylbiacridine has been shown to proceed via two consecutive one-electron oxidations, the first step occurring at a more anodic potential than the second.
THECHEMILUMINESCENCE OF LUCIGENIN (dimethylbiacridinium ion) (L2+) was first reported in 1935 by Gleu and Petsch ( I ) .
They observed intense chemiluminescence when lucigenin was treated with hydrogen peroxide in basic solution. It has been shown that this reaction is of a redox type (2, 3) but the exact nature of the redox steps has not yet been ascertained. Totter (3) has briefly discussed the electrochemical reduction of lucigenin at a DME in aqueous solution, and has shown it to be pH independent. To date, no further work on the electrochemistry of lucigenin has been reported. A detailed study of the electrochemical behavior of lucigenin is necessary for analysis of the redox steps of the chemiluminescent reaction. Such studies are also of interest in explaining the electrochemical generation of lucigenin chemiluminescence, reported by Legg and Hercules (4).
Present address, Department of Chemistry, California State College at Long Beach, Long Beach, Calif. a Present address, Department of Chemistry, Muhlenberg College, Allentown, Pa. 3 Please address all correspondenceto this author.
(1 j K. Gleu and W. Petsch, Augew. Chem.,48,57 (1935). (2) J. R. Totter, Photochem. Photobiol., 3,231 (1964). (3) J. R. Totter and G. E. Philbrook, ibid., 5,177 (1966). (4) K. D. Legg and D. M. Hercules, J. Amer. Clzem. SOC.,91,1902 (1969).
1650
ANALYTICAL CHEMISTRY, VOL. 44, NO. 9, AUGUST 1972
In the present study, we have undertaken an investigation of the electrochemical behavior of lucigenin in aqueous and nonaqueous systems. The reduction of lucigenin has been shown to be a two-step process: first a one-electron charge transfer step yielding the mono-cation radical of lucigenin (L. +); second, a rapid disproportionation of the radical to dimethylbiacridine (DBA) and lucigenin. The oxidation of DBA has been shown to occur via two consecutive one-electron oxidations yielding lucigenin as the product.
5.0
-
4 9 w
c C
3.0
-
2.0
-
.o
-
-1.0
-
I
I
-
4.0
: 0.0 u
0
0
-2.0
-
-
-3.0
+I
I
Figure 1. Current-voltage curves for lucigenin and oxygen in DMSO at Pt (Tetrabutyl ammonium perchlorate supporting electrolyte, 0.1 M). Cell resistance 1000 ohms, scan rate 3.33 x V/sec - Lucigenin --.-Oxygen
(I++)
EXPERIMENTAL Reagents. Lucigenin (nitrate salt) was obtained from Columbia Organic Chemicals and was recrystallized twice from 1 : 1 methanol-ethanol before use. Dimethylbiacridine was prepared according to the method of Decker and Petsch (3, and was purified by recrystallization from chloroform. Tetrabutyl ammonium perchlorate was prepared from 1M tetrabutyl ammonium hydroxide (Southwestern Analytical) by addition of 35% perchloric acid until the solution was acidic (pH -2). The precipitated perchlorate was collected by filtration and washed several times with hot water to remove all traces of acid. The perchlorate was then remethanol and dried crystallized twice from 70z water-30 in cacuo at 100 "C. All other chemicals were reagent grade and were used without further purification. Solvents. Distilled water was used in the preparation of all aqueous solutions. Dimethyl sulfoxide (Matheson, Coleman and Bell, spectroscopic grade) was dried over molecular sieves before use. Solutions. Solutions of DBA were prepared by saturating the solvent with DBA and measuring the concentration spectrophotometrically at 420 nm (emax at 420 nm for DBA is 1.66 X 1041itermole-' cm-I). Apparatus. A potentiostat constructed by D. W. Shive (6) was used in the electrochemical studies. This instrument has sufficiently fast response time for the cyclic scan voltamrnetry used in this study. The current integrator for coulometric measurements consisted of a Nexus SQ-3 operational amplifier with a 10 pf capacitor in the feedback loop. A standard three-electrode cell was used in all electrochemical measurements. The reference electrode, Ag/AgCl in 0.1M aqueous KCI, was separated from the bulk of the solution with a porous Vycor plug. The counter electrode consisted of a 2 cm2 platinum foil isolated from the test solution by a fine glass frit. A platinum sphere of 5.53 x 10-2 cm2 area served as the indicator electrode for nonaqueous studies and a Hg electrode in aqueous systems. Tetrabutyl ammonium perchlorate, O.IM, was used as supporting electrolyte in nonaqueous solutions and 0.1M KCl in aqueous media. For coulometric measurements, a 2 cm2 platinum foil working electrode was used. Provision for deoxygenation with nitrogen was provided. Procedures. Standard techniques were used to obtain current-voltage curves for scan rates up to 0.333 V/sec. ( 5 ) H. Decker and W. Petsch, J. Prtrkr. Chem., 43,211 (1935). (6) D. W. Shive, Ph.D. Thesis, M.I.T., Cambridge, Mass., 1969.
For scan rates greater than 0.33 V/sec, a Wavetek Model 112 function generator was used to drive the potentiostat and the current-voltage curves were recorded on an oscilloscope. Graphical corrections (7) for charging currents were made in order to obtain peak currents from these data. Coulometry was used to determine the number of electrons involved in the electrochemical reduction of lucigenin, n,, and in the oxidation of DBA, n,. Known concentrations of lucigenin in DMSO were introduced into the electrochemical cell and the peak current, ipc,was determined. After electrolysis at -0.35 V, the new concentration of L2+ was determined from a graph of ipc cs. concentration, The procedure was repeated and n was taken as the average of several runs. A Varian Model E-3 ESR instrument was used to detect the presence of radicals in the lucigenin reduction. A twoelectrode cell was used consisting of a platinum wire indicator electrode and a mercury drop reference. RESULTS Cyclic Voltammetry. Electrochemical studies were done in nonaqueous media. Typical current-voltage curves for lucigenin and oxygen in dimethylsulfoxide (DMSO) at a platinum electrode are shown in Figure 1. The solid line is the cyclic voltammogram obtained with only lucigenin present in the solution. The wave peaking at -0.30 V us. Ag/AgCl in 0.1M KC1 corresponds to the reduction of lucigenin and that peaking at $0.45 V corresponds to the oxidation of lucigenin's reduction product. The dotted line in Figure 1 is obtained with only oxygen present. The wave peaking at -0.9 V is the reduction of oxygen to superoxide (Oz-) and the anodic wave peaking at - 0.85 V is the reoxidation of superoxide to oxygen. These curves indicate that oxygen reduction would not interfere with the kinetics of the L2+reduction in DMSO. DBA has been shown ( 4 ) to be the electrochemical reduction product of lucigenin nonaqueous solvents. Referring to Figure 1, if the potential was swept anodically from the open circuit potential with only lucigenin present, no anodic peak was observed at +0.45 V. However, if lucigenin were reduced, the peak at +0.45 V appeared. When only DBA was present in solution, no wave peaking at - 0.30 V was observed until an anodic scan had been made over the DBA oxidation wave at f0.45 V. (7) R. S.Nicholson and I. Shain, ANAL.CHEM., 36,706 (1964).
ANALYTICAL CHEMISTRY, VOL. 44, NO. 9, AUGUST 1972
1651
Table I. Scan Rate Study of Lucigenin Reduction in DMSO 2.14 X 10-3M Lucigenin ip/Cov1’2 log Ep -Ep/2 v, V/sec i,(mA) x C,iv (mV) 1.67 x 10-3 0.93 1.063 0.11 40 3.33 x 10-3 40 1.28 1.037 -0.19 1.94 0.993 -0.59 40 8.34 x 10-3 1.67 x 2.61 0.944 -0.89 45 3.33 x 10-2 40 3.52 0.901 -1.19 0.841 -1.59 8.34 x 10-2 5.20 40 1.67 x lo-’ 40 6.95 0.795 -1.89 9.25 3.33 x 10-1 0.749 -2.19 45 4.64 x 10-1 10.4 0.713 -2.34 50 1.17 14.6 0.630 -2.73 60 60 2.32 18.4 0.564 -3.04 4.64 24.1 0.523 -3.34 60 17.7 65 41.2 0.457 -3.92 50.8 0.449 -4.12 70 28.0 56.0 72.0 0.450 -4.42 70 140 113 80 0.446 -4.82 1.07 X 10-2MLucigenin 1.67 x 10-3 5.60 1.280 0.81 36 1.300 3.33 x 10-3 8.00 0.51 42 1.228 0.11 40 12.0 8.34 x 10-3 1.063 1.67 x 14.7 42 -0.19 3 . 3 3 x 10-2 1.070 46 -0.49 20.9 0,990 30.6 44 -0.89 8.34 x 56 -1.19 38.3 1.67 x 10-1 0,876 3.33 x 10-1 0.850 60 -1.49 52.5 56 0.805 -1.64 58.7 4.64 x 10-1 0.719 64 -2.05 1.22 85.0 0.605 64 -2.47 3.16 115 0.569 90 -3.00 10.8 200 0.417 34.0 260 100 -3.50 0.327 160 -4.04 380 118
ill-defined values of iPa due to merging of the oxidation wave for DBA with the solvent oxidation. Studies of fast scan-rates in solutions of lucigenin more dilute than 10-3M were difficult to interpret because of the large capacitive currents encountered. As the limit of usable scan rates and lucigenin concentrations (50-100 V/sec with 1 0 3M lucigenin) was approached, an anodic peak was observed at -0.2 volt us. Ag/AgCl. A typical cyclic voltammogram exhibiting this peak is shown in Figure 3. ESR. A two-electrode cell consisting of a mercury reference and platinum wire indicator in a thin glass tube was used in an ESR cavity in a n attempt to detect the presence of radicals during the reduction of lucigenin in DMSO. A steady state potential of -0.40 V was applied. A broad ESR signal was detected indicating low concentrations of some radical, but the nature of this radical could not be determined. DISCUSSION
Nicholson and Shain (7) have presented a theoretical discussion of electrochemical behavior at a stationary electrode, as applied to cyclic scan methods. Since the peak current varies with the square root of the scan rate (v), peak currents can be normalized for scan rate by dividing them by v1’2. A plot of normalized peak current (current function) cs. scan rate gives a horizontal straight line for either a reversible or irreversible charge transfer step, the reversible line being lower than that of the irreversible case (7). If, however, there are any coupled chemical reactions which affect the electrochemistry, a plot of normalized peak current us. - scan rate will divide from the horizontal. Referring to Figure 2, one sees a decreasing current function for the Table 11. Scan Rate Study of DBA Oxidation in DMSO reduction of lucigenin as the scan is increased. This indicates Y, V/sec ip,mA i,/V”Z x 10-1 a chemical reaction coupled with the electrochemical reaction 1.67 x 10-1 10.3 3.15 at these scan rates. Table I1 indicates that the current func8.34 x 8.8 3.05 tion for the oxidation of DBA does not vary with scan rate; 3.33 x 10-2 5.6 3.06 therefore, there are no chemical reactions coupled with the 1.43 x 3.8 3.20 charge transfer step. 6.67 x 10-3 2.5 3.07 Since no pH dependence was observed for the lucigenin 2.33 x 10-3 1.5 3.10 reduction, a preceding or succeeding reaction of lucigenin or its reduction intermediates or products with either protons or hydroxyl ions is unlikely. Lucigenin is stable in the soluIn nonaqueous media, the peak potentials for lucigenin tions studied showing no other detectable tautomeric forms. reduction and DBA oxidation were found to be unaffected Thus, the possibility of any chemical reaction preceding the by saturating the solution with either gaseous HC1 or solid charge transfer step is highly unlikely. The possibility of a KOH, indicating the process to be pH independent. chemical reaction after the charge transfer step to account Coulometry. An average of 8 coulometric measurements for the observed deviations of ip, (Figure 2) is the only of the number of electrons involved in lucigenin reduction alternative. in DMSO yielded n, = 1.96 i 0.14. Coulometry was also ESR studies indicated evidence of radicals during the used to determine the value of nu for reoxidation of DBA in reduction of lucigenin. Janzen et al. (8) have reported a DMSO. The results of 3 coulometric determinations for the radical produced during the reduction of lucigenin and DBA oxidation yielded an average of nu = 2.06 =t0.16. identified it as being the mono-cation radical of lucigenin, Scan Rate Studies. To determine what, if any, kinetic L.+. This radical was observed when a constant potential complications were involved in the electrochemistry of luciwas applied corresponding to the reduction potential of genin, studies were performed on the reduction wave of lucigenin. The L . + radicals observed were on the very lucigenin and the oxidation wave of DBA as a function of limit of detectability and it was estimated that their steadyscan rate. state concentration was no greater than 10PM (9). The results of these studies on the lucigenin reduction wave This ESR evidence for L . + in solution, and the electroin DMSO are shown in Table I. The current function i,/Cov1’2, chemical indication of a chemical reaction coupled with the normalized at low scan rates, is plotted us. scan rate in Figure 2 along with similar data for lucigenin concentrations of 1.07 X (8) E. G. Janzen, J. B. Pickett, J. W. Happ, and W. DeAngelis, 10-2M and 2.14 X 10-*M. Similar studies were performed J . Org. Chem., 35, 88 (1970). On the oxidation wave of DBA in DMSO and are Presented (9) E. G. Janzen, University of Georgia, private communication, in Table 11. Scan rates greater than 2 X 10-1 V/sec gave 1970. 1652
0
ANALYTICAL CHEMISTRY, VOL. 44, NO. 9, AUGUST 1972
charge transfer step, justify two possible mechanisms for the reduction of lucigenin:
+ e- e L.+ (charge transfer) L . + + S. ff: DBA + S+ (chemical oxidation) L2f
(1)
where S is either solvent or supporting electrolyte Lz+ K
+ e-
2L. + i L2f
L . + (charge transfer)
(11)
+ DBA (Disproportionation).
Both reactions involve a one-electron charge transfer, followed by a chemical reaction of the mono-cation radical DBA as the product. Mechanism I has an overall n of 1 while mechanism I1 has an n value of 2. Since the measured value of n, is 2, mechanism I1 is favored. The nature of the effect of a following chemical reaction on the observed electrochemical behavior depends on the time scale of the electrochemical measurements and the value of the rate constant of the chemical reaction (10). In the case of mechanism 11, two extreme cases may be considered. First, at slow scan rates the time scale of the electrochemical measurement is large compared with that of the chemical reaction, Le., the chemical reaction is complete. For mechanism I1 then, one should observe behavior corresponding to a two-electron, irreversible reduction. Second, at fast scan rates the time of the electrochemical measurement is small compared to that of the chemical reaction. In this case, no chemical reaction occurs during the time of measurement. For mechanism I1 in this instance, one should observe behavior corresponding to a one-electron reduction with no kinetic complications. Nicholson and Shain (7) showed that the current function for a twoelectron irreversible reaction is 2.2 times the current function for a one-electron reversible reaction. Referring to the data of Table I and Figure 2, one may see that the current function in the irreversible part of the curve (low scan rates) is 2.38 times the current function in the reversible region. Since the current function at low scan rates may be high due to convection in solution, this value corresponds well with the theory of Nicholson and Shain. Studies at scan rates between the first and second cases will be greatly influenced by the following chemical step. The value of the scan rate, such that the following reaction is negligible, is dependent on the value of the rate constant of the following reaction. The larger the rate constant, the higher the scan rate must be to negate the effect of the chemical reaction. The decrease in normalized peak current with increasing scan rate (Figure 2) is indicative of a following chemical reaction. Several aspects of this curve are important. The curve obtained with 2.14 X 10-3M1~cigeninshows indications of becoming horizontal at scan rates greater than 20 V/sec. As the concentration of lucigenin was increased, the leveling out of the curve became less apparent. It was also noted that at scan rates greater than 20 V/sec in the 2.14 X 10-3M lucigenin solution, a new anodic wave appeared as seen in Figure 3. At low scan rates, the normalized peak current also tended to level off. The leveling of the current function at high scan rates, as explained above, is indicative of approaching the pure chargetransfer step of the reduction, with the following chemical (10) R. N. Adams, “Electrochemistry at Solid Electrodes,” Marcel Dekker, New York, N.Y., 1969.
10-3
10-2
IO-’
I
IO
loo
1000
v ( V/sec) Figure 2. Current function (normalized at low scan rates) scan rate for lucigenin reduction in DMSO 0-0-0 A-A-A
c-n--0
CS.
2.14 X 10-3Mlucigenin 1.07 X 10-2Mlucigenin 2.14 X 10-*M lucigenin
c
c
2 ‘ i (3 .-V
U 0
f
3 c
C
L
q
,
J
a+Iv
ov
- IV
0
c +
‘doitage vs ACJIA~CL Figure 3. Current-voltage curve obtained for 10-3M lucigenin in DMSO. Scan rate 80 Visec reaction not affecting the measurement. The appearance of an anodic wave at the scan rates where this leveling occurs supports this premise. This anodic wave corresponds to the oxidation of the intermediate which is involved in the chemical reaction. The net effect of lowering the scan rate due to the IR drop in the cell may be seen in Figure 2. At scan rates greater than 10 V/sec, the normalized peak currents are lower as the lucigenin concentration is increased. Thus, the values of v1I2 employed in normalizing the peak currents are higher than those encountered in reality. The overall effect on the scan rate dependence is to broaden the curve. The value of E , - EP/*is a measure of the peak shape of the reduction process (IO). At low scan rates (see Table I),
ANALYTICAL CHEMISTRY, VOL. 44, NO. 9, AUGUST 1972
1653
LO@
1,
Figure 4. Current function Qp us. kinetic parameter hd, after normalizing hd with a secondorder rate constant k,' = 4.41 X 104M-' sec -I
+++
Theoretical curve (ZI) Experimental points
the value of Ep - E,iz for the lucigenin reduction is 40 mV. At high scan rates this approaches a value of 80 mV. A completely reversible electrochemical reaction shows an Ep - E,,z of 57 mV (7, IO). At high scan rates, where one approaches the condition of a charge transfer step with no kinetic complications, the difference of 65 mV indicates that true reversibility has not been reached. This is also borne out by the large (0.3V) separation of the lucigenin reduction peak potential and that of the anodic peak appearing at high scan rates. From the above interpretation of the scan rate studies, coupled with the measured value of n, and chemically feasible reactions, it is evident that there is a chemical reaction following charge transfer. The most probable reaction is that shown in mechanism 11, an initial one-electron reduction to the mono-cation radical (L .+) and disproportionation of L to lucigenin (L") and DBA. The current function plots indicate a one-electron charge transfer followed by a chemical reaction producing a net two-electron, irreversible reduction which is in agreement with the measured value of n,. The appearance of an anodic peak at fast scan rates, where the following chemical step has a negligible effect, corresponds to the oxidation of L.+. The exact region of scan rates at which the charge transfer step only is observed is questionable because of the effect of IR drop on the scan rate. Nicholson and Shain (7) have treated only first-order or pseudo-first-order following chemical reactions. However, an analogy may be drawn between an irreversible, following chemical step (Nicholson and Shain, Case VI) and the proposed disproportionation. Nicholson and S h a h (7) give a value of krja (for case VI reactions) of 0.1 for the onset of seeing the anodic peak, corresponding to the oxidation of the electrochemical intermediate ; k r is the rate constant of the following reaction and a = nFv/RT, where n is the number of electrons involved in the charge-transfer step, and v is the scan rate in V/sec. If one assumes, from the data of Table I and Figure 2, that the onset of the anodic wave occurs at an actual scan rate of 10 Vjsec, the first-order rate constant is then calculated to be 1.7 X 101 sec-1. If the following chemical 9
1654
+
step is second-order, its rate must equal that of the first-order reaction. Assuming that the concentration of L.$- at the electrode surface is equal to the bulk concentration of lucigenin (2.14 X 10-3M) and calling the second-order rate constant k , one calculates: k,(L*+) = kf.(L.+)z or ky.
=
8 X 103M-' sec-I
This calculated value of k, is only an approximation indicating the minimum possible value of kr because the maximum value for the concentrations of L . + has been assumed. Mastragostino et al. (11) have provided a theoretical treatment of the specific case of a following disproportionation step. They note that the peak current variation is larger than a direct proportion to the concentration, as is seen in Table I and Figure 2. The rate constant, k , for the disproportionation reaction can be obtained by plotting the current function q P = kip/Cov1'2us. the kinetic parameter RT C, Ad = kr - -. Normalization of experimental results nf v with the theoretical curve (11) will yield the value of k,. The data of Table I were treated in this manner and a plot of \ZIP us. Ad along with the theoretical curve appears in Figure 4. A good fit of experimental results with the theory is obtained. The rate constant, kf, which gives this agreement is 4.41 X 104M-1 sec-1. The deviation of the experimental points from the theoretical curve at low scan rates (large log Ad) can be explained, once again, by convection interfering with pure diffusion and yielding higher values of the current function. Plots of Ep - EPi2(normalized us. Ad) also agree with the theory of Mastragostino et al. (11), from log Ad = - 1 to 3. Therefore, the wave shape for the reduction of L2+ is also consistent with the proposed mechanism. (11) M. Mastragostino, L. Nadjo, and J. M. Saveant, Electrochim. Actn, 13, 721 (1968).
ANALYTICAL CHEMISTRY, VOL. 44, NO. 9, AUGUST 1972
Olmstead and Nicholson (12) have indicated the feasibility of evaluating k l by constructing working curves for the ratio of anodic to cathodic peak current U S . kf COT. Here C" is the bulk concentration of L2+and 7 is the time from E" to the switching potential, Ex. Since the appearance of the anodic wave occurred in a region extremely high capacitive current (Figure 4), it is not possible to utilize this diagnostic criterion to evaluate k Nevertheless, the appearance of the anodic peak at a value of log Ad = -1.58 substantiates the proposed mechanism and is consistent with the estimated rate constant. The scan rate data on Table I1 for DBA oxidation show no dependence of normalized peak current with scan rate. This is indicative of an electrochemical oxidation uncomplicated by chemical reactions. The experimental value of n, = 2 indicated a net two-electron oxidation. Figures 1 and 3 show no reversible cathodic wave in conjunction with the anodic DBA wave, indicating the two-electron oxidation of DBA to be a net irreversible process. Since true twoelectron oxidations in organic systems are unlikely, it is more probable that the oxidation passes through a one-electron oxidation intermediate. The one-electron oxidation product of DBA is L.'. By reference to Figure 4, it may be seen that L.+ itself will be oxidized to lucigenin at the anodic potential necessary to oxidize DBA to L . + . Thus, the oxidation of DBA appears to fit the scheme: (12) M. L. Olmstead and R. S . Nicholson, ANAL. CHEM.,41, 862 (1969).
DBA
- e-
--t
L e + - e- + L2+
or two, consecutive one-electron oxidations occurring at the same potential. In summary, the electrochemical behavior of lucigenin in DMSO appears to best fit the following scheme : L*+
2L.+ -% L2+ DBA
+ e- e L . +
+ DBA for the reduction of lucigenin
- e- e L . + -
e-
-+
L2+ for the oxidation of DBA
The value of k , has been estimated as 4.41 X 104M-l sec-l. Potential step methods by Booman (13) for following disproportionation reactions may yield a more refined estimate of k,. ACKNOWLEDGMENT
We thank David Hume and Leon Klatt for reading the manuscript and for their helpful suggestions.
RECEIVED for review February 1, 1972. Accepted March 30, 1972. This work was supported in part through funds provided by the U S . Atomic Energy Commission under Contract AT(30-1)905. K.D.L. was a Predoctoral Fellow, 1965-68. (13) G. L. Booman and D. T. Pence, ibid., 37,1366 (1965).
Ion-Selective Electrode Study of Copper(1) Cornplexes in Acetonitrile L. F. Heerman and G . A. Rechnitz Department of Chemistry, State University of New York, Buffalo, N . Y . 14214
A cuprous sulfide-membrane ion-selective electrode was used for the potentiometric measurement of copper(l) ion concentration in acetonitrile with tetraethylammonium perchlorate or sodium perchlorate as supporting electrolytes. The electrode showed an almost Nernstian behavior (slope 55-56 mV/decade) for copper(l) ion concentrations down to lo-5M in pure solutions and to at least 3 x l W 7 M in the presence of complexing ligands. The electrode was used for the measurement of the stepwise formation constants of copper(l) complexes with the halides in acetonitrile. Log p1 and log p2 values were as follows: CI-, 4.9 and 10.7; Br-, 3.8 and 7.8; and I-, 3.2 and 6.4. The electrode might be particularly useful for the study of interactions of copper(1) ions with organic ligands in acetonitrile. As an example, complex formation between copper(1) and thiourea has been investigated. The experimental data suggest Cu[S=C(NH2)&+ as the predominant complex species for which the formation constant has been evaluated as log p2 = 6.3. COPPER(I)IONS are stable in a number of organic solvents, e.g., acetonitrile ( I ) and nitromethane (2). The stability of copper(1) ions in such solvents has been explained by an (1) I. M. Kolthoff and J. F. Coetzee, J. Amer. Chem. SOC.,79, 1852 (1957). (2) I. V. Nelson, R . C. Larson, and R. T. Iwamoto, J . Inorg. Nucl. Chem.,22,279 (1961).
increase of the solvation energy of copper(1) ions and/or a decrease of the solvation energy of copper(I1) ions with respect to the solvation energies of these ions in water ( 2 ) . The formation constants of copper(1) complexes with chloride in acetonitrile have been measured by polarographic methods (3). The formation constants of copper(1) complexes with the halides and thiocyanate have been determined by potentiometric measurements using a copper amalgam indicator electrode (4). In this work, the formation constants of copper(1) complexes with the halides and with thiourea in acetonitrile have been measured using a novel cuprous sulfide-membrane ionselective electrode as the indicator electrode. Ion-selective electrodes have been used for a large number of complex formation studies in aqueous solutions (5, 6). Furthermore, ion-selective electrodes have been used for potentiometric titrations in a number of organic solvents, e.g., acetonitrile, and their application for the quantitative (3) S. E. Manahan and R. T. Iwamoto, Inorg. Chem., 4, 1409 (1965). (4) J. K. Senne and B. Kratochvil, ANAL.CHEM., 43,79(1971). ( 5 ) J. N. Butler, "Ion-Selective Electrodes," R. A. Durst, Ed., NBS Special Publication No. 314, U.S. Government Printing Office, Washington, D. C. 1969,p 143. (6) G. A. Rechnitz, Accounts Chem. Res., 3,69 (1970).
ANALYTICAL CHEMISTRY, VOL. 44, NO. 9, AUGUST 1972
1655
