lonlc Aggregation of the Solvated Electron
the measurement of redox potentials correlating them with spin density distribution in radicals and radiosensitization efficiencies.3z Correlation between two-electron redox potentials and any parameter involving one-electron transfer, such as rates of electron transfer or the energy of the first unoccupied orbital, reflects in effect the correlation between the two- and the one-electron transfer if such a correlation exists.
References and Notes (1) Supported in part by the U.S. Atomic Energy Commission. (2)(a) L. Michaelis, J. Biol. Chem.. 96, 703 (1932):(b) B. Elema. ibid., 100, 149 (1933). (3)W. M. Clark, "Oxidation-Reduction Potentials of Organic Systems". Williams and Wilkins, Baltimore, Md., 1960. (4)L. Edsberg, D.Eichlin, and J. Gavis, Anal. Chem., 25, 798 (1953). (5) I. M. Kolthoff and T. B. Reddy, J. Nectrochem. SOC.. 108, 980 (1961). (6) M. E. Peover. J. Chem. SOC.,4540 (1962). (7)S. Wawzonek, R. Berkey, W. Blaka, and E. Runner, J. Nectrochem. SOC., 103, 456 (1956). (8) E. Mueller and W. Dilger, Chem. Ber., 106, 1643 (1973). (9)M. E. Peover and B. S. White, Electrochim. Acta, 11, 1061 (1966). (10) M. E. Peover, Na?ure(London), 193, 475 (1962). (11)(a) J. Lilie. G. Beck, and A. Henglein, Ber. Bunsenges. Phys. Chem., 75.
1509 458 (1971);(b) M. Gratzel and A. Henglein, ibid., 77, 2 (1973),and the three following papers. (12)P. S. Rao and E. Hayon, Biochem. Biophys. Res. Commun., 51, 468 (1973). (13) P. S.Rao and E. Hayon, J. Am. Chem. SOC.,96, 1287,1295 (1974). (14)K. B. Patel and R. L. Willson. J. Chem. Soc.,faraday Trans. 1, 69, 814 (1973). (15)P.S.RaoandE. Hayon. J.Phys. Chem., 77, 2753(1973). (16)P. S.Raoand E. Hayon, Nature (London). 243, 344 (1973). (17)P. W. Preisler, E. S. Hili. R. G. Loeffel. and P. A. Shaffer, J. Am. Cbern. Soc., 81, 1991 (1959). (18)P. A. Shaffer and P.W. Preisler. lnd. Eng. Chern. News, 11, 236 (1933). (19)B. Venkatoraman. B. G. Segal, and G. K. Fraenkel. J. Chem. Phys., 30, 1006 (1959). (20)G. Vincow and G. K. Fraenkel, J. Chem. Phys.. 34. 1333 (1961). (21) L. Michaelis, M. P. Schubert, R . K. Reber, J. A. Kuck, and S. Granick, J. Am. Chem. SOC., 60, 1678 (1938). (22)Y. I. Ibn, D. Meisel, and G. Czapski, lsr. J. Chem., 12, 891 (1974). (23)J. Chevalet, F. Rouelle. L. Gierst. and J. P. Lambert, Electroanal. Chem. lnterface Electrochem.. 39, 201 (1972). (24)R. L. Willson. Chem. Commun.. 1249 (1971). (25)P. S.Rao and E. Hayon, J. Phys. Chem., 77, 2274 (1973). (26)J. H. Baxendale and H. R. Hardy, Trans. faraday SOC.,49, 1140 (1953). (27)J. H. Baxendale and H. R. Hardy, Trans. Faraday SOC.,49, 1433 (1953). (28)D.Behar. G.Czapski, J. Rabani, L. M. Dorfman, and H. A. Schwartz, J. Phys. Chem., 74, 3209 (1970). (29)L. K. Patterson and J. Liiie. lnt. J. Radiat. Phys. Chem., 6,129 (1974). (30) K. Eiben and R . W. Fessenden, J. Phys. Chem., 75, 1186 (1971). (31)L. E. Bennett, Prog. horgan. Chem., 18, 27 (1973). (32)D.Meisel and P. Neta, submitted to J. Am. Chern. SOC.
Ionic Aggregation of the Solvated Electron with Lithium Cation in Tetrahydrofuran Solution' Bradley Bockrath and Leon M. Dorfman' Department of Chemistry, The Ohio State University, Columbus, Ohio 432 10 (Received September 9. 1974; Revised Manuscript Received April 30, 1975) Publication costs assisted by the Energy Research and Development Administration
Aggregation of the solvated electron with lithium cation in tetrahydrofuran forms an ion pair which has been investigated by the pulse radiolysis method. The ion pair (Li+,e,-) has a near-infrared absorption 1180 nm and a molar extinction coefficient of 2.28 X lo4 A4-I cm-' a t the maximum. Absoband with A,, lute rate constants have been determined for reactions of this species with anthracene, biphenyl, and dibenzylmercury. These reactivities are compared with the reactivities of (Na+,e,-) and e,- in T H F solution.
Introduction Aggregation of the solvated electron with alkali metal cations forms a variety of species in a number of weakly polar liquids. The sodium cation-electron pair, (Na+, e,-), has been observed in d i g l ~ m eand ~ ~in~ t e t r a h y d r ~ f u r a n . ~ The species Na-, formed by reaction of two solvated electrons with sodium cation, has been observed in ethylenediamine5-7 and in t e t r a h y d r o f ~ r a n . The ~ , ~ ion pairs (K+,e,-) and (&+,e,-) in tetrahydrofuran have recently been reported.IOOptical absorption spectra have been determined. The k i n e t i c ~ ~ > ~ofJ ~both J ' the formation of such species and the reactions they undergo have been studied. In some solvents, the optical absorption spectrum of the solvated electron is very greatly altered upon ion pair formation. Thus, the absorption maximum for (Na+,e,-) in T H F is 890 nm4 compared with a maximum a t 2120 nm for e,- in THF,'zJ3 corresponding to a change of 0.8 eV in the transition energy. Rate constants for the attachment of the
electron to various substrates are substantially diminished4J1upon ion pair formation with sodium cation. The reactions involved in this pairing e,- t Na'
es-
(Nat,es-) eNa-
(1)
are by no means unique to sodium cation. To obtain a broader knowledge of alkali metal-electron pairing, information has been obtained about the pairing of the solvated electron with lithium cation in THF. The results have been obtained by pulse radiolysis of T H F solutions of various dissociative lithium salts. We report here the optical absorption spectrum of the species (Li+,e,-) as well as rate constants for the attachment of this species to anthracene and to biphenyl. Experimental Section The source of the electron pulse, as in our earlier studies,14 was a Varian V-7715A electron linear accelerator, deThe Journal of Physical Chemistry, Wol. 79, No. 15, 1975
B. Bockrath and L. M. Dorfman
1510
livering 3-4-MeV electrons a t a pulse current of about 300 mA for pulse duration of 100-1500 nsec and about 600 mA for pulse duration less bhan 80 nsec. Electron pulses of 200to 800-nsec duration were used in this work. The transient optical absorptions in the region from 600 to 1100 nm were observed using an RCA 7102 photomultiplier which has S-1 spectral response. Our infrared detector,13 a solid state diode employing a diffused junction of indium antimonide, manufactured by Barnes Engineering Co., was used for observation in the region from 900 to 2000 nm. The 10-90% rise time of the electronic detection systems is 80 nsec for the infrared detector and considerably less for the RCA 7102 photomultiplier. A Bausch and Lomb grating monochromator, Type 33-86-25, fh.5 was used. Corning filters were selected to elimicate second-order components from the analyzing light beam. Our standard reaction cells,14 with high-purity silica windows and a cell length of 20.0 mm, were used with a double pass of the analyzing light beam. The T H F was purified first by refluxing under argon, for several hours, a solution containing benzophenone and excess sodium metal. The solvent was then distilled through a glass-bead-packed column, the middle fraction being retained. It was then degassed and vacuum distilled into a storage bulb containing a mirror of freshly distilled potassium. Solvent was vacuum distilled from this bulb into the reaction cells just prior to the runs. Pulse radiolysis experiments with this pure solvent show only the spectrum of the solvated electron i t ~ e l f . ~ ~ J ~ Anthracene and biphenyl (Aldrich) were zone refined, with a nominal purity of at least 99.9%. Dibenzylmercury (Alfa Inorganics) was recrystallized from ethanol, dried under vacuum, and stored in the dark until used. Lithium perchlorate (Alpha Products, 99.5%) was found, in our analysis for potassium by atomic absorption, to contain only 0.001% K by weight. Lithium bromide (Matheson Coleman and Bell, purity nominally >99%) was found, in OUT assay, to contain 0.01% K by weight. Lithium chloride (hialter, purity nominally 99.8%) contains, according to the manufacturer's assay, 0.005% K and 0.002% Na. These salts were used as supplied after thorough drying under vacuum. Lithium tetraphenylboron was prepared from sodium tetraphenylboron (Fisher reagent grade) and lithium chloride by cation exchange.15 The freshly prepared salt was recrystallized three times by addition of cyclohexane to a dichloroethane solution. Solutions of lithium tetraphenylboron become discolored when exposed to the atmosphere. Precautions were taken to reduce contact with the atmosphere by conducting the preparation and recrystallization in a glove bag filled with argon and storing the salt under vacuum until immediately before weighing on a pan balance and transfer to the reaction cell. Results a n d Discussion Optical Absorption Spectrum of (Li+,e,-). The transient absorption band which we assign to the lithium cation--solvated electron ion pair, (Li+,es-), was obtained by measuring the optical density, a t different wavelengths, immediately after an electron pulse in T H F solutions of several dissociative lithium salts. The data are shown in Figure 1. The data combined in Figure 1 were obtained in separate experiments using either LiCl, LiBr, or LiC104 as the added salt. The salt concentration was sufficiently high so that the ion pairing reaction, e,Li+, was essentially complete at the end of the electron pulse. This was readily
+
The Journal of Physical Chemistry, Vol. 79, No. 15, 1975
01
I
I
10
I
I
1
20
15 Frequency (cm-' x (0-31
Figure 1. Spectrum of (Li+.eg-) in THF at 25' obtained in solutions of: 0 , LiClOa (0.035 F): 0,LiCl (0.064 F); and +, LiBr (0.12 F). The absorption maximum is at 1180 nm with a molar extinction coefficient of 22,800 at the maximum and width at half-height of 6270 cm-'. The line is calculated using a gaussian curve on the low-energy side of the maximum and a lorentzian curve on the high-energy side, and is not a best fit for the experimental points.
verified by observations in the near-infrared which showed that e,- is not present after the pulse because of this effective scavenging. It may be readily seen that, when the optical densities obtained from the separate solutions of the three salts are normalized to the absorption maximum, the data define a common absorption band with the maximum a t 1180 nm. Since the same band is obtained from the three lithium salts it is evident that the anion does not play a role in the formation of this transient which is indicated to be (Li+,es-). It should be noted that in pure THF12J3 only the band corresponding to the solvated electron, with maximum a t 2120 nm, is seen. In short, the "blank" for these runs contains no hint of a band a t 1180 nm which is thus produced solely by addition of the various lithium salts. Moreover, an atomic absorption analysis of a T H F solution of LiC104 (0.035 F ) showed a potassium content of only 0.04 ppm, consistent with the salt concentration, and indicating no potassium contamination from unknown sources. Further scavenger experiments indicate that this transient is composed of both the lithium cation and a solvated electron. When dibenzylmercury was used to scavenge the band a t 1180 nm, formed in lithium perchlorate solutions, a new band with a maximum a t 330 nm appeared. This new species is identified as benzyllithium since the absorption band corresponds to that reported16 for PhCH*-Li+. Under our experimental conditions, this band could have arisen only by the reaction: (Lif,es-) + (PhCHz)2Hg
-
PhCH,-Li*
+
PhCHzHg (2)
This conclusion is well founded for several reasons. First, we have previously established that, in irradiated T H F solutions of dibenzylmercury, the solvated electron is the precursor of free benzyl anion." Secondly, the absorption spectrum of PhCH2-IAi+ is readily distinguished from that of PhCH2- or that of PhCH2- paired with other alkali metal cations,18 since the absorption maximum differs in the case of sodium cation and presumably also for potassium cation. Finally it was shown that no significant amount of %- could react directly with dibenzylmercury. To ensure that the solvated electron reacted only with lithium cation, the solution was made with lithium perchlorate (0.028 F ) in rather large excess over dibenzylmercury (2.5 X M). Observation of the band a t 1180 nm under these conditions shows only decay, as expected. Thus we may confidently
Ionic Aggregation of the Solvated Electron
1511
TABLE I: Optical Absorption Spectrum of the Solvated Electron and Its Lithium Ion Pair i n THF Solution -
Trail sition maximum, energy, nm eV
Absorption
Species
2120" 1180 89Qb 112v 1400'
Width at Shift from e,", eV
half
-
height, cm-'
3450 6270 1.39 7000 W*, es-) 1.10 6000 (CS+, ea-) 0.88 5000 a Data taken from ref 12. 6 Data taken from ref 4. Data taken e, (Lit, esq) (Na'. es-)
0.58 1.04
0.46 0.81 0.52 0.30
from ref I O .
rule out formation of benzyllithium by the reaction sequence e,(PhCH&Hg followed by BhCH2- i- Li+. Reaction 2 thus seems well established, providing assur1180 nm, ance that the benzyllithium precursor, with,,,A is in fact (Li+,e,-). Table I contains a comparison of the optical transition energies for the absorption bands of the solvated electron itself, the lithium-electron pair, the sodium-electron pair, and recent data for the potassium-electron pair and cesium-electron pair,1° all in THF. The change in transition energy for the solvated electron induced by pairing with lithium ion is considerably less than that induced by pairing with sodium ion. If this shift in transition energy is taken as an indication of the strength of the coupling in the ion pair, the relative effect for Li+ and Na+ is difficult to rationalize only on the basis of the ionic radii of the bare alkali metal cations. If ionic radius was the dominant parameter in determining the change in transition energy, the smaller lithium ion might be expected to induce the greater shift. Our contrary observation suggests that solvation of the alkali metal cation is the important phenomenon which determines the interionic distance in the pair; in short, we are dealing with a solvent-separated pair. Any meaningful evaluation of the role of the solvent would require knowledge of the number and the geometric arrangement of solvent molecules in a t least the first solvation shell. Such detailed microscopic information is lacking. Attention has, however, been drawn to the role of the solvent with these ions on the basis of other macroscopic properties. The limiting conductance15J9 of Li+ in THF solution was found to be lower15 than that of Na+, a fact attributed to the relative size of the solvation shell of T H F molecules bound to the metal cations. Of the two ions, Li+ apparently has the larger solvation shell. The effect on the spectrum, which we observe and which we interpret as stronger coupling of the electron by sodium cation, may be rationalized similarly on the basis that Li+ is solvated to a greater degree than Na+ in the ion pair, (M+,e-), just as for the free ion in solution. A greater degree of solvation, together with the lesser effect of effective nuclear charge for Li+ than for Na+, would reduce the coulombic attraction between the cation and the electron in the pair. The shape of the absorption band of (Li+,e,-) conforms reasonably well to the same shape function as does the band for e,- and for (Na+,e,-), namely, a gaussian curve on the low-energy side and a lorentzian curve on the high-energy side of the maximum. The fit of the data in the region to the right of the band maximum in Figure 1 does not appear to be quite as close as for other cases. The total hand-
+
width at half-height, for the five spectra in THF, is also compared in Table I. All the metal cation-paired species have a considerably broader band, on a wave number scale, than does es-. The molar extinction coefficient of (Li+,e,-) a t the band maximum was found to be 2.28 x IO4 M-l. This value was determined, as before: by using aromatic compounds to scavenge the transient species: (Li*,es-)
+ arerie
-+-
Li+,arene"
(3)
The extent of growth or decay, after the electron pulse, of optical density a t any given wavelength, depends upon the relative values of the extinction Coefficient of (Li+,e,-) and of Li+, arene.'-. At wavelengths for which the extinction coefficients of these two species are equal, neither growth nor decay is observed. With anthracene as the scavenger, this wavelength of equivalence was found to be 720 nm, which happens to correspond to an absorption peak o f anthracenide ion. With biphenyl as scavenger the wavelength of equivalent extinction coefficients was found to be 676 nm. The extinction coefficient of (Li+,e,-) was then obtained from the two sets of data by using the known extinction coefficients of sodium anthracenide20 and sodium biphenylide20 in THF, 10,000 and 12,500 M-l cm-l a t the absorption peak, respectively. Good agreement was obtained, the average value being 2.28 X lo4 M-l cm-l. The oscillator strength of the 1180-nm band was determined from the equation
f = 4.32
X
10-9(1.065W1~,Gf 16511H'1,2L)6,,,u
(4)
where W1/zG and W1/zL are the portions of the half-width on the gaussian and lorentzian side, respectively. This gives a value o f f = 0.90 for the band of (Li+,e,-), which i s similar to the value found for other one-electron species, namely, es- in THF13 (f = 0.85) and (Na+,e,-) in THF4 (f = 1.01, as well as for e,- in a variety of liquids. The optical absorption data thus indicate that ion pairing of e,- in T H F induces a shift in the absorption to higher transition energies. The magnitude of the shift depends upon the specific cation involved, being greater for sodium than for lithium. The absorption band for es- is considerably broadened by cation pairing, and the extinction coefficient reduced, while the oscillator strength remains roughly the same. Reaction Kinetics of (Li+,e,-). Table I1 contains absolute rate constants for elementary reactions of (Li+,e,-) with three different substrates. Two are examples of electron attachment to aromatic compounds, reaction 3, and one of a dissociative electron attachment, reaction 2. Values of the rate constants for analogous reactions of e,and of (Na+,e,-) have been included for comparison. These rate constants were determined in each case by observing the decay in the absorption a t 1000 nm in the absence as well as in the presence of appropriate concentrations of the added reactants. This wavelength is near the absorption maximum of (Li+,e,-), but sufficiently far removed from the absorptions arising from the products (either the corresponding aromatic radical anions or benzyllithium) so that there is no spectral overlap. In the absence of added substrate, the decay of (Li+,e,-) in LiC104 solutions (0.03 to 0.06 F ) followed a first-order rate law with a half-life of about 3 psec. In the presence of anthracene (0.46 and 1.0 X M ) and of dibenzylmercury (1.3 and 2.4 X low4M ) the decay was again first order with n halfThe Journal of Physical Chemktry, Vol. 79,No. 15, 1975
E. Bockrath and L. M.Dorfman
1512
TABLE 11: Absolute Rate Constants for Reactions of (Li+ ,es-),(Na+,es- ), and es- in THF Solution
k(phCH2)*Hg is less than unity for (M+,e,-), while it is 4 for es-.
at 25" ( M - I sec-I X
Reactant
(Lit, e s - )
(Na', ea-)
Anthracene 2.65 1.00 0,55a Biphenyl 1.8 0.79* Dibenzylmercury a Data taken from ref 4. Data taken from ref 17.
e, -
ll.oa
Acknowledgment, We are indebted to Mr. Ed Ray for maintaining the electron linear accelerator and the electronic detection equipment. We are grateful to Dr. James Gavlas for performing the atomic absorption analyses.
2.P
References and Notes life reduced to less than 300 nsec. The rate constant for the reaction was then evaluated from the pseudo-first-order constant and each substrate concentration. The rate constants in Table I1 are the average of the two determinations at the indicated substrate concentrations. The separate values agree to within f15%. The uncertainty of the rate constant values in Table I1 is f20%. In the case of biphenyl, k 3 was obtained in an experiment using lithium tetraphenylboron as the dissociative salt. These attachment rate constants for (Li+,e,-) are all close to the diffusion-controlled limit judging from the approximate value for kdiff calculated from the Smoluchowski equation.21s22
where q1 the viscosity, isl9 4.61 X loW3P, and r , and r b are the interaction radii, which are taken, as an approximation, to be equal. We thus estimate kdiff = 1.4 X 1O1O A4-l sec-l, quite comparable to the experimental values. Two properties of the rate constants for the metal catim-coupled species, compared with those for ea-, are noteworthy. First, the values for (M+,e,-) are somewhat lower. Secondly, there is a difference in selectivity; the ratio kphz/
The Journal of Physical Chemlstry, Vol. 79, No. 15, 1975
(1) This work was supported by the US. Atomic Energy Commission under Contract No. AT(ll-1)-1763. (2) J. G. Kloosterboer, L. J. Giling, R. P. H. Rettschnick, and J. D. W. VanVoorst, Chem. Phys. Lett., 8, 462 (1971). (3) G. Ramme, M. Fisher, S. Claesson, and M. Szwarc, Proc. R. SOC., Ser. A, 327, 467 (1972). (4) B. Bockrath and L. M. Dorfman, J. Phys. Chem., 77, 1002 (1973). (5) S. Matalon, S. Golden, and M. Ottolenghi, J. Phys. Chem., 73, 3098 (1969). M. G. DeBacker and J. L. Dye, J. Phys. Chem., 75,3092 (1971). J. L. Dye, M. G. DeBacker, J. A. Eyre, and L. M. Dorfman, J. Phys. Chem., 76, 839 (1972). G. A. Salmon and W. A. Seddon Chem. Phys. Lett., 24,366 (1974). M . T. Lok, F. J. Tehan, and J. L. Dye, J. Phys. Chem., 76, 2975 (1972). G. A. Salmon, W. A. Seddon, and J. W. Fletcher, Can. J. Chem., 52, 3259 (1974). M. Fisher, G. Ramme, S. Claesson, and M. Szwarc, Proc. R. SOC.,Ser. A, 327, 481 (1972). L. M. Dorfman, F. Y. Jou, and R. Wageman, Ber. Bunsenges. Phys. Chem., 75, 681 (1971). F. Y. Jou and L. M. Dorfman, J. Chem. Phys., 58, 4715 (1973). W. D. Fellx, B. L. Gall, and L. M. Dorfman, J. Phys. Chem., 71, 384 (1967). D. N. Bhattacharyya, C. L. Lee, J. Smid, and M. Szwarc, J. Phys. Chem., 69, 608 (1965). R. Waack and M. A. Doran, J. Am. Chem. SOC.,85, 1651 (1963). B. Bockrath and L. M. Dorfman, J. Am. Chem. SOC., 96, 5708 (1974). B.Bockrath and L. M. Dorfman, J. Am. Chem. SOC., in press. C. Carvajal, K. J. Tolle, J. Smid, and M. Szwarc, J. Am. Chem. SOC., 87, 5548 (1965). J. Jagur-Grodzinski, M. Feld, S. L. Yang, and M. Szwarc, J. Phys. Chem., 69, 628 (1965). M. Smoluchowski, Z.Phys. Chem., 92, 129 (1917). P . Debye, Trans. Electrochem. SOC.,62, 265 (1942).
