J. Phys. Chem. 1990, 94, 7682-7683
7682
Does Isotopic Substltution Affect the Reduction Potential of Aromatic Moiecules? Timothy T. Goodnow and Angel E. Kaifer* Department of Chemistry, University of Miami, Coral Gables, Florida 33124 (Received: January 24, 1990; In Final Form: May 17, 1990)
Reduction potentials for anthracene, pyrene, perylene, and nitrobenzene, their perdeuterated forms, and ISN-enrichednitrobenzene were measured by using voltammetric techniques. The results indicated that the perdeuterated forms have lower solution electron affinities than the isotopically unmodified counterparts since the former exhibit slightly more negative reduction potentials. However, 15N enrichment was not observed to cause any significant effects under our experimental conditions.
Introduction Starting in 1986, Stevenson and co-workers have reported several electron exchange equilibrium constants involving a molecule and an isotopic isomer.' The generalized form of the equilibrium is
+
*A A*A- 4- A (1) where *A represents a form (isotopic isomer) enriched in a heavy isotope (D, 13C, IsN, etc.). These authors have found equilibrium constant ( K ) values that differ substantially from unity.' The findings for I3C- and I5N-labeled compounds are controversial since they appear to contradict well-established theoretical predictiom2 Differences in solution electron affinities between A and *A, as indicated by K values different from 1, should be reflected in measurable differences between the one-electron reduction potentials of A ( E O ) and *A ( E o * ) since K can be expressed as K = exp(-AGO/RT)
(2)
AGO = -F(EO* - E O 1
(3)
where Reduction potentials can be quickly and conveniently determined by using voltammetric techniques. We report here easily verifiable experimental data obtained using cyclic voltammetry (CV) and differential pulse voltammetry (DPV) that appear to support the findings of Stevenson and cc-workers on the deuterated compounds.
Experimental Section Materials. Anthracene (AN), pyrene (PY), perylene (PR), and nitrobenzene (NB) were obtained from Aldrich and used without further purification. The perdeuterated materials were purchased from MSD isotopes (PY-dlo, PR-d12, NB-d5) and Aldrich (AN-d,,,). I5N-enriched nitrobenzene was obtained from MSD isotopes and used as received. Acetonitrile (HPLC quality, EMS Science) was stored over molecular sieves and handled under nitrogen. Tetrabutylammonium hexafluorophosphate (Fluka, electrochemical grade) was used as received. Equipment. Cyclic voltammetry was performed with a Princeton Applied Research (PAR) Model 175 universal programmer, a Model 173 potentiostat/galvanostat,and a Model 178 digital coulometer equipped with circuitry for IR compensation. Voltammograms were recorded on a Houston Model 2000 X-Y (1) (a) Stevenson, G. R.; Espe, M. P.; Reiter, R. C. J . Am. Chem. Soc. 1986,!08,532. (b) Stevenson, G. R.; ESP, M. P.; Reiter, R. C. J. Am. Chem. Soc. 1986,108,5760. (c) Stevenson, G. R.; ape,M.P.; Reiter, R. C.; Lovett, D. J. Nature 1986,323, 522. (d) Stevenson, G. R.; Reiter, R. C.; Espe,M. P.; Bartmess, J. E. J . Am. Chem. Soc. 1987, 109, 3847. (e) Stevenson, G. R.; Reiter, R. C.; Au-Yeuno,W.; Pescatore, J. A.; Stevenson, R. D. J. Org. Chem. 1987.52, 5064. (0 Lauricella, T. L.; Pescatore, J. A.; Reiter, R. C.; Stevenson, R. D.; Stevenson, G. R. J. fhys. Chem. 1988, 92, 3687. (g) Stevenson, G. R.; Sturgeon, B. E.;Vines, K.S.;Peters, S.J. J . Phys. Chem. 1988, 92,6850. (h) Stevenson, G. R., personal communication. (2) Marx, D.; Kleinhesselink, D.; Wolfsberg, M. J . Am. Chem. Soc. 1989, 111, 1493.
0022-3654/90/2094-7682$02.50/0
TABLE I: Cyclic Voltammetric Data on the Reduction at 23 O C of Several Aromatic Hvdrocarbons and Their Isotonic Isomers no. of no. of 4 / 2 : range of std compd samples msrmts mV E l J 2values dev
anthracene anthracene-d,,
2 2
pyrene
I I
pyrene-dio perylene perylene-dI2 nitrobenzene nitrobenzene-d5 nitrobenzene-ISN
2
5 6 3 3 5
I
3
2 1 1
6 3 3
-2403 -2418 -2536 -2551 -2123 -2131 -1554 -1 567 -1555
-2409, -2420, -2543, -2555, -2126, -2134, -1558, -1579, -1558,
-2391 -2416 -2532 -2545 -2121 -2129 -1549 -1556 -1551
7 2 6 6 2 3 4 12 4
values are given in millivolts against the potential of the '*l' ferricinium couple in the same solution. ferrocene '
'
'
recorder. Differential pulse voltammetry was carried out with a Bioanalytical Systems Model 100 electrochemical analyzer connected to a Houston plotter. Procedures. Solutions for electrochemical analyses were freshly prepared, deoxygenated, and kept under nitrogen for the duration of each individual experiment. The solvent system was dry acetonitrile containing 0.1 M tetrabutylammonium hexafluorophosphate as the supporting electrolyte. A platinum flag and a glassy carbon disk were used as the auxiliary and working electrodes, respectively. The working electrode was polished with 0.05-pm alumina-water slurry on a felt surface, sonicated in distilled water, and air dried prior to use. An aqueous sodium chloride saturated calomel half-cell was used as the reference electrode. However, to avoid the irreproducibilities associated with liquid junctions in potential measurements, ferrocene was also added to the test solutions, and potentials are reported vs the potential of the ferrocene/ferricinium couple unless stated otherwise. Neutral alumina was always added to the test solution and stirred before the voltammetric measurements to remove residual water from the system. All measurements were performed at room temperature (23 f 1 "C). Results If A and *A have the same reduction potential, eqs 2 and 3 show that the equilibrium constant K must be exactly 1. Any difference in the value of the reduction potentials translates into K values different from 1. Therefore, the experimental determination of reduction potentials for A/A- redox couples provides a simple alternative methodology to obtain thermodynamic values for equilibria of type 1, Voltammetric techniques afford a number of fast and convenient methods for the determination of reduction potentials. Because of their widespread use and availability, we have selected here CV and DPV to determine the reduction potential corresponding to the process A + e- e A- for several aromatic molecules and their perdeuterated forms. CV affords reduction and oxidation peaks separated by 57 mV for a reversible A/A- couple. The average of these two peak potentials yields the half-wave potential ( E l l 2 ) ,which can be identified with the thermodynamic reduction potential if the oxidized (A) and reduced (A-) species are assumed to have the same 0 1990 American Chemical Society
The Journal of Physical Chemistry, Vol. 94, No. 19, 1990 7683
Reduction Potential of Aromatic Molecules
TABLE II: AGO Values for the Process A- + *A + A A AN PY PR NB NB
*A AN-dlo PY-dlo PR-dIz NB-dS "N-NB
AGO," cal/mol (this work) 350 f 168 350 196 180 f 83 300 f 292 20 f 130
+ *A-
AGO: cal/mol (literature) 410 340 310 93
I
-1.500
I
c
OObtained by using eq 3, at 23 OC. *Reference 1; data obtained at -100 OC except NB values obtained at -65 OC. cSee ref Id. This value was found to be very sensitive to medium effects.
diffusion coefficients, a common and well-accepted assumption in electrochemical measurement^.^ Table I shows the half-wave potentials obtained in our CV experiments for the reversible reduction of AN, PY, PR, NB, and their corresponding perdeuterated forms as well as I5N-enriched NB. In all cases the voltammograms showed peak-to-peak potential differences equal or quite close to the theoretical value, indicating that the A/Aredox couples behave reversibly. Thus, the half-wave potentials obtained in these experiments can be taken as very good approximations to the thermodynamic reduction potentials. It should be stressed that the potentials in Table I are free from liquid junction effects since they were measured against the potential of the ferrocene/ferrocinium couple in the same test solution. The standard deviation values shown in the table were obtained from repeated measurements. All perdeuterated compounds exhibit reduction potentials slightly more negative than the corresponding isotopically unmodified compounds. However the largest difference observed is 15 mV, which is still rather small. To ensure that the observed differences are meaningful, we compared the reduction potential of the deuterated and isotopically unmodified compounds using t tests. This statistical procedure reveals that the null hypothesis is not valid, i.e., the values obtained are different at the 95% confidence level in the AN, PY, and PR cases. The NB vs N B 4 and NB vs I5N-NBcomparisons were not so clear because the series of values obtained for NB did not conform to a normal distribution. Table I1 gives the AGO values (23 "C) obtained from our reduction potential data with eq 3 and the values reported by Stevenson and co-workers using ESR techniques.I Both sets of results indicate that the deuterated materials have a lower solution affinity than their isotopically unmodified counterparts. The values show qualitative agreement. It is interesting to point out that Stevenson et al. generated the anion radicals A- in T H F through reduction with potassium metal.la*b This procedure results in K+-anion radical ion pairs. Under our experimental conditions, the anion radicals are generated in a medium that contains only tetrabutylammonium cations, and no ion pairing is expected. Also the electrochemical data presented here were obtained at room temperature, while Stevenson's data were recorded at lower temperatures. These differences do not introduce much variation in the AN, PY, and PR results because of the extensive charge delocalization throughout the molecular structure of the corresponding anion radicals and the small entropic changes in electron-transfer processes between similar aromatic molecule^.^ However, the NB results seem to be more sensitive to solvent and ion effects due to the larger extent of charge localization in the nitro group.Id This is perhaps the main reason behind the discrepancies observed between our data and Stevenson's for NB and their isotopic isomers. To verify our CV measurements, we selected AN and AN-dlo, as the case yielding the largest difference in reduction potentials, and performed differential pulse voltammetry on solutions of these two compounds. As expected, we obtained DPVs showing re(3) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; Wiley: New York,1980 Chapters 5 and 6. (4) Alper, J. S.;Silky, R. J . Chem. Phys. 1970, 52, 569.
v
-2 0
+
4 -2.x)o
EIWLT)
Figure 1. First derivative of the differential pulse voltammetric response recorded on a glassy carbon electrode immersid in a 0.1 M TBA+PF{/CH$N solution also containing (top) 1.O mM AN and (bottom) 1.O mM AN-dlo. The DPVs were obtained by using a pulse amplitude of 25 mV and a scan rate of 5 mV/s.
versible behavior for both the AN/AN- and AN-dlo/AN-dl{ redox couples. The perdeuterated material exhibited a peak potential at -2005 mV vs SSCE while the undeuterated material was at -1990 mV vs SSCE. The excitation waveform for the DPV measurements consisted of a 25-mV pulse amplitude sequence. The half-wave potential is therefore 12.5 mV greater than the observed peak potential. Half-wave potentials of -1993 mV for AN-dloand -1977 mV for AN vs SSCE were calculated. The observed difference of 15 mV was maintained when the potentials were measured against the ferrocene/ferrocinium couple. Therefore, the DPV results also show that AN-dlois reduced at a more negative potential than AN. This is clearly evident in Figure 1, which shows the first derivative of both DPVs. In these plots the intercept of the first derivative curve with the potential axis indicates the relative position of the peak in the original DPV. The comparison of both curves in Figure 1 clearly demonstrates that the deuterated material (Figure 1, bottom) exhibits a more negative peak potential, and thus, a more negative reduction potential than the isotopically unmodified material (Figure 1, top). Furthermore, these DPV results agree very well with the previously described CV data. In conclusion, we have shown here that voltammetric techniques provide an independent means for the evaluation of free energy differences for electron-exchange processes between isotopic isomers in solution. Our experimental data reveal a definite reduction potential difference between perdeuterated and isotopically unmodified aromatic hydrocarbons in the cases of AN, PY, and PR. This is in good qualitative agreement with Stevenson's results. We cannot be as definitive with the heavier isotope substitutions (such as I3C and 15N)because within the accuracy limits of our experiments we find no 15Nisotope effects in NB. Even for NB/NB-d5 the agreement of our data with Stevenson's is poor. These discrepancies may perhaps be attributed to the variations in solvent and ionic composition resulting from the different experimental conditions.
