The Photochemical Behavior of Cobalt Complexes ... - ACS Publications

(1). We have undertaken photochemical investigations of cobalt(II1) complexes of the type tr~ns-Co~~'(trans- ..... pH 3. 0 Ion pair. h [I-] inferred f...
0 downloads 0 Views 889KB Size
2223

PHOTOCHEMICAL BEHAVIOR OF COBALT COMPLEXES However, if it is assumed that the trap density of the M , as seems to be present glass is approximately , ~ ~tunneling ~~ must the case for 10 M NaOH g l a s ~such take place over distances of the order of 25 A in order to explain a major fraction of the decay of e,-. Hence, if the estimate of trap density is of the correct order of magnitude, this mechanism is probably not important here. (3) The stabilized electron may move as such. That is, the electron trap with the electron in it may be randomly moving by dielectric relaxation of the medium. This possibility is supported by the sharp rise in the decay rate of e,- which is observed in the temperature region where the glass begins to soften. As already noted, it is of course possible that the diffusion of e,- occurs by a combination of these mechanisms. Thus, it is conceivable that diffusion over relatively large distances, say >10 8, occurs primarily by mechanism 3, and that the dominance of this mechanism increases with temperature, while short-range displacements may take place by tunneling processes, primarily the tunneling from trap to scavenger, which should be relatively more important at low temperatures.

Such a combination of diffusion mechanisms may be one reason for the time-dependent reaction rate of e,as shown in Figure 5. Thus, it seems possible that the relatively fast initial decay is associated mainly with tunneling cf part of e,- directly into the scavenger molecule, i e . , that part of e,- which is trapped sufficiently close to scavenger molecules, while the slower decay componept is due to diffusion of e,- as such, ie., by mechanism 3 above. This hypothesis is indeed supported by the observation that the temperature dependence of these two decay components appears to be different.g A further study of the reaction kinetics of e,- in glasses may possibly shed some light on the more general problem of the time dependence of the reaction of eaq- in its early stages ( i e . , in very concentrated solutions) which so far seems to be based primarily on theoretical arguments.

Acknowledgment. We are indebted t o Professor Aa. Ore and Mrs. J. Mossige for constructive criticism of the manuscript. (18) B. G. Ershov and A. K. Pikaev, Rad&. Res. Rev., 2 , 1 (1969).

The Photochemical Behavior of Cobalt Complexes Containing Macrocyclic (NJ Ligands. Oxidation-Reduction Chemistry of Dihalogen Radical Anions' by Sally D.Malone and John F, Endicott* Department of ChemktTu, Wayne State Undversity, Detroit, Michigan 48202

(Received February 18, 197%')

Publication wets asshted by the Public Health Service

The electron transfer reactions of dihalogen radical anions, XZ-, have been examined with cobalt(II1)and with cobslt(I1) substrates. The XZ- oxidations of Co(trans [14]diene)2+all proceed at nearly diffusion-controlled rates. The observation of this reaction for 12- substantiates the prediction that 1 9 - is a stronger oxidant than IZor 18-. No direct evidence for XZ- reduction of cobalt(II1) complexes has been obtained. Indirect evidence is presented that 1 2 - reduces Co(NH&12+.

Introduction Following the classical study of Grossweiner and Mathesonz the spectral and kinetic properties of the dihalogen radical anions, IP-,~ Brz-,S84 and C12- 3--5 have been reasonably well characterized.s These radicals have often been postulated as likely intermediates in one-equivalent reactions of various metal ions with aqueous solutions of the halogen^^-^ or halides.I0 Despite their ease of generation and detection, and de-

spite their apparent mechanistic importance, there have been few direct observations of the reactions of (1) Partial support of this research by the Public Health Service (Grant AM 14341) is gratefully acknowledged. (2) L. I. Grossweiner and M. S. Matheson, J . Phys. Chem., 61, 1089 (1957). (3) M. 9. Matheson, W. A. Mulac, and J. Rabani, ibid., 69, 2613 (1965). (4) M. E. Langmuir and E. Hayon, ibid., 71, 3308 (1967). (5) M. Anbar and J. K. Thomas, ibid., 68, 3829 (1964). The Journal of Physical Chemistry, Vola76, No. 16,197%'

2224

SALLY D. MALONEAND JOHN F. ENDICOTT

dihalide radical anions with transition metal subs t r a t e ~ . ~ ~There ~ ~ - ' ~has been some direct evidence for theX2- oxidations of transition metal substrates. 11-13 It is to be observed, however, that the irreversibility of their redox disproportionation reactions2-6 (1) suggests that these species may in principle also function as reducing agents.

pared in an inert atrno~phere,'~ and the compound staoredin stoppered containers under nitrogen until use. B . Apparatus. The instrumentation used to generate and record flash photolysis data has been described," as have the various continuous photolysis apparatuses. 25--27 Absorption spectra were measured using a Cary 14 recording spectrophotometer. The 20-cm cells used for flash photolysis were made with 2x2xz 2 x (1) quartz windows and jacketed with a quartz tube concentric with the sample ~ell.~'8~* The flash pulse had We have undertaken photochemical investigations a 50-psec detectable duration for 250-J pulses.l1 about of cobalt(II1) complexes of the type tr~ns-Co~~'(transFor less intense pulses the pulse duration was shorter, [l4]diene)X2l4in order to examine some aspects of the becoming about 15 psec for a 50-5 pulse. kinetics and energetics of recombination reactions inC . Procedures. To eliminate oxygen from solutions volving dihalogen radical anions and Co"(trans [ 141to be photolyzed, flasks containing the weighed amount diene) 15-17 species. Since the tetradentate macrocyclic of solid to be dissolved were evacuated then filled with trans [l4]diene ligand remains coordinated to cobalt (11) Cr2+-scrubbed nitrogen. The appropriate amount of for long periods in aqueous s o l ~ t i o nradical , ~ ~ ~ recom~~ previously deaerated solvent was quickly added by bination reactions are expected to be an extremely impipet, and the solutions were then transferred via pipet or syringe and serum cap into previously N2 flushed systems. Samples for spectral measurements were II I syringed into preflushed 1-cm silica cells with round CHZ-N HN-CH, tops onto which serum caps were fixed. Each solution I

-

CH2-NH I

+

I

N-CHS II

trans[ 14 Niene

portant component of the charge transfer to metal (CTTM) photochemistry of the trans-CoI1I(trans[14]diene)X2 species. Furthermore CoII(trans[ 14ldiene) species are known to be weakly reducing16-18 and low spin," so these are good model systems in which t o examine the reactivity of low spin (3d7) cobalt(I1) complexes with simple radicals. I n addition to the X2- oxidations of Co1Ir(trans[14ldiene), we have looked for evidence of Xz- reductions of cobalt(II1) complexes.

Experimental Section A. Materials. Reagent grade acids and simple salts were used without further purification. Distilled water, redistilled from permanganate in Pyrex apparatus, was used for all solutions. For solution deaeration, tank nitrogen was passed through a chromous-perchloric acid scrubber to remove the last traces of oxygen. The C O ( N H & ~ + ,C~O ~ ( N H ~ ) ~ B and ~ ~ + Co,~~ (NH3)512+ 2o complexes were prepared as described in the literature and metathesized to the perchlorate salts by recrystallizing from dilute perchloric acid when desired. Preparations of (trans[l4]diene). 2HC104 21 and N ~ ~ [ C O ( C O ~ ) ~ ]used . ~ Hto~ prepare O , ~ ~ [Co(trans[14]diene)C03]C104and subsequently [Co(trans[l4]diene)XZlC101 or Co(trans [14]diene)Xs were followed as previously d e ~ c r i b e d . ~The ~ ! ~preparation ~ of [Co(trans[ 14ldiene) (O2CCHs)OH]C1O4has been described elsewhere.16 The [Co(trans[14]diene) ](C104)2 was preThe Journal of Physical Chemistry, Vol. 76, No. 16,1979

(6) For recent reviews see: (a) A. Treinin, "Radical Ions," E. T. Kaiser and L. Kevan. Ed.. Interscience. New York. N. Y.. 1968. Chapter 12, pp 544-551; (b) J. K. Thomas, Adv. Radiatwn Chem., 1, 103 (1969). (7) P. R. Carter and N. Davidson, J . Phys. Chem., 56, 877 (1952). (8) J. H. Crabtree and W. P. Schaeffer, Inorg. Chem., 5 , 1348 (1966). (9) W. H. Woodruff, D. C. Weatherburn, and D. W. Margerum, ibid., 10, 2102 (1971). (10) A. J. Fudge and K. W. Sykes, J . Chem. SOC.,119 (1952). (11) G.Caspari, R. G . Hughes, J. F. Endicott, and M. Z. Hoffman, J . Amer. Chem. Soc., 92, 6801 (1970). (12) A. T. Thornton and A. 5. Lawrence, Chem. Commun., 443 (1970). (13) (a) T. L. Kelly and J. F. Endicott, J. Amer. Chem. SOC.,92, 5733 (1970); (b) ibid., 94, 1797 (1972). (14) tram [14]diene = 5,7,7,12,14,14-hexamethyl-1,4,8,1 l-tetraazacyclotetradeca-4,ll-diene. (15) The axial coordination positions of the cobalt(I1) complex are labile, and there is some limited evidence16.1' that a t equilibrium there are no ligands strongly bonded in the axial positions in these complexes. (16) M.P.Liteplo and J, F. Endicott, Inorg. Chem., 10,1420 (1971). (17) D. P. Rillema and J. F. Endicott, J. Amer. Chem. Soc., submitted for publication and unpublished observations. (18) D. P. Rillema, J. F. Endioott, and E. Papaconstantinoe, Inorg. Chem., 10, 1739 (1971). (19) W. C. Fernelius, Inorg. Syn., 2, 217 (1940); H. S. Booth, ibid., 1, 186 (1939). (20) R. G.Yalman, Inorg. Chem., 1, 16 (1962). (21) N.F. Curtis and R. W. Hay, Chem. Commun., 524 (1966). (22) H. F.Bauer and W. C. Drinkard, J . Amer. Chem. SOL,82, 5031 (1960). (23) N. Sadasivan and J. F. Endicott, ibid., 88, 6468 (1966). (24) N.Sadasivan, J. A . Kernohan, and J. F. Endicott, Inorg. Chem., 6, 770 (1967). (25) J. F. Endicott and M. Z. Hoffman, J . Amer. Chem. Soc., 87, 3348 (1965). (26) J. F.Endicott, M.Z. Hoffman, and L. S. Beres, J . Phys. Chem., 74, 1021 (1970). (27) W. L. Wells and J. F. Endicott, ibid., 75, 3075 (1971). (28) See ref 13.

PHOTOCHEMICAL BEHAVIOROF COBALTCOMPLEXES from which kinetic measurements were taken was flashed only once and then discarded. I n some cases we have used cut-off filter solutions13t o prevent irradiation of deep ultraviolet absorption bands or to minimize the amount of decomposition during a single flash. Flash photolysis data were recorded in the usual manner. ' l After converting observed transmittance values into absorbance values and correcting for substrate absorbances, data were treated in either first-order (log (A - A,) us. time) or second-order (1/A us. time) manner. The order of reaction was determined from plots of log A. us. tl/,.ll

Results A . Continuous Photolysis. The 254-nm irradiation of solutions of Co(trans [14]diene)(OH~)2~+ and Co(trans[ 14]diene)C12+ results in the formation of Co(tr~ns[l4]diene)~+.Overall quantum yields are esti(in 0.1 M HC104) and mated as (3.5 =t 1.2) X (3 f 1) X (in 0.1 M HC104, 0.9 M KCl), respectively, for the diaquo and dichloro complexes. The determination of primary quantum yields in these systems is extremely difficult for a t least three reasons. (1) As reported in detail below Co(trans[14]diene)2+ is oxidized efficiently by many of the primary and secondary radicals and by oxidized ligand species (e.g., X2) formed from photoreduction of Co"'(trans[14]diene)Xz. (2) Co(trun~[l4]diene)~+ is very readily oxidized by Oz especially at high halide concentrations. (3) The absorption maxima of the Co(tran~[14]diene)~+ product) (€825 = 2.6 X lo3; 8260 = 1 X lo3) and original substrates (e.g., for Co(trans[14]diene)Clz+, €825 = 1.6 X lo3; 6250 = 1.4 X lo4) are not sufficiently different that the yields of products can be readily determined for small extents of photolysis. To attempt a better estimate of the primary quantum yield we have also photolyzed Co(trans [14]diene)C12+ in pure methanol. Under these conditions one expects reactions 2 and 34g11 Co(trans[14]diene)Clz+

+ hv +

+ + C1C1* + CH3OH +HC1 + *CH20H Co(trun~[14]diene)~+ C1-

(2) (3)

in competition with (4) and followed by (5)ze through (7). Since these experiments in methanol were per-

c1. + c1- JT Clz-

+ Co(truns[l4]diene)Cl2+ Co(tran~[l4]diene)~+ + 2C1- + CH20 + H + Clz- + Co(tr~ns[l4]diene)~+ -+.CH20H

(4)

-+

(5)

Co(trans [14]diene)Cl~+ (6)

+

c1* Clz-

Clz

+ c1-

(7) formed at very low [Cl-1, (4), (6), and (7) can be assumed to be relatively unimportant. We have not -+

2225 observed ( 5 ) , but assume that it may account for half the observed cobalt(II).29 The overall quantum yield of Co(trun~[14]diene)~+ observed under these conditions was (3 f 1) X lo2which implies a primary yield of about (1.5 f 1) X The 254-nm irradiation of Co(trans 114ldiene)( 0 2 CCHs)OH+ in neutral solution or in Na02CCH3solution (0.1 or 0.01 M ) led to the formation of Co(trans[14]diene)2+. Although we were unable to purge sample solutions of C 0 2 before photolysis (the hydrolysis of this cobalt(II1) complex is acid catalyzed), we did find that CH4 and CZH, were not detectable as photolysis products. We estimate the quantum yield to be (0.12 f for formation of Co(tr~ns[14]diene)~+ 0.03) for this complex. When [Co(trans [14]diene)(Na)n]C104was irradiated a t 254 nm, the absorbance changes a t 560 nm were consistent with ~Co(lrans[14]diene)S+ = 0.04. B . Flash Photolysis Studies. Dihalide Radical Anion Decay Rates. To be able t o compare values of the X2- decay rate constants (ICl) appropriate to our experimental conditions we have generated C12-, Br2-, and 12-from several sources. These results are summarized in Table I. Values of kl in Table I have 1520% precision except in the cases noted. Dihalide Radical Anion Reactions with Co(trans[14]diene)z+. As noted in Table I, transients generated from Co(trans[14]diene)Clz+ and Co(trans[l4]diene)Brz+ do not exhibit a simple second-order decay over more than one reaction half-life. The observed curvature of the second-order plots is consistent with a competition between the natural decay mode (1) and reaction 8 with the cobalt(I1) fragment which formed in the photoreduction of Co(trans[14]diene)Xz-. 15,80 To ob-

Y

+ Co(tr~ns[14]diene)~+ + Xz- +

+

C0~~I(trans[l4]diene)XY X-

(8)

tain reliable rates of reaction 8 we have examined the pseudo-first-order decay rates of X2- in the presence of variable amounts of Co(tr~ns[14]diene)~+.From a plot of the pseudo-first-order decay constants of Xzvs. [Co(tr~ns[l4]diene)~+] we have obtained the values of kain Table 11. With regard to the data in Table I1 it should be noted that Co(trans[14]diene)I2+ is unstable (in excess I-) with respect to 13-and Co(trans[14]diene)2+.1E~18~B1 Since Co(trans [14]diene)2+has appreciable absorptivity in the near-ultraviolet, we have used solutions of the highly absorbing ion pair (CO(NH&~+,1-)32,33as our (29) D.Meyerstein, private communication. (30) The nature of the axial ligands in the immediate cobalt(II1) products is not known. (31) W. Latimer, "Oxidation Potentials," Prentice-Hall, Englewood Cliffs, N.J., 1952. (32) M. G. Evans and G . H. Nancollas, Trans. Faraday SOC.,49, 363 (1949). (33) M.T.Beck, Coord. C h m . Rev., 3, 91 (1968). The Journal of Physical Chemistry, Vol. 76, N o . 16,1978

SALLY D. MALONEAND JOHN F. ENDICOTT

2226

Table I : Dihalide Radical Anion Decay Rate Constantsa

1x-1,

Specie8 irradiated

C1- c , d Co(truns[14]diene)Cl~+4 Br- c , d Co(NH&Br2 + Co(trans[141diene)Brz + e Co(tetraen)Brz+

I- C

J

Co(NHS),’+, Ieo( ”3)6IZ +

x

Ionio strength

ki X 10-0, M-8 sea-1

1.2 1 4.1 1-3 1-10 4 2-5 9 1

1.1 1.1 1.1 0.20 0.20 0.20 0 * 020 0.020

8.4 8.2 d~ 4 5.8 3.2 6.4 f 0.8 3.0

0.10

3.5 5.4

7

0.12

i

1x2-It-o

M

M

1.0 1.0 1.0 0.20 0.10 0.20 0.010 0.010 1.0 x 10-3 (2 x 10-6)h

108,b

7

a pH 1 except as indicated. Mean values of kl are quoted. Average deviations were f 2 0 % except as indicated. * Extrapolated concentration. A range of [XZ-] 1-0 is indicated in cases that several experiments gave differing values. c Potassium salt. d Solutions saturated with NtO. Second-order plots generally were significantly curved; kl estimated from initial slope. pH 3. 0 Ion pair. h [I-] inferred from [Iz-]~~.o, Ks = 8 X lo4 M-’and the amount of Co(NHa)612+decomposed in the flash. i At very low [I-] the decay of 1%- becomes first order; see ret 2. The first-order rate constant wm 2.8 X l o 4 sec-1.

Table 11: Rate Constants for Oxidations of Co(trans[l4]diene)2+ by Dihalide Radical Anions

XI

-

Clz Brz-

Source of Xa-

Co(trans[141diene)Clz + Co(trans[l4]diene)Br~+ Co(NH&BrZ CO(NH&~+,I- c +

I2 Q

pH 1; halide in parentheses.

E-],

Range of added

M”

Ionic strength

[Co(trana[14]diene)a+I, Mb

ks X 10-9, M-1 seo-1

1.0 (Cl-) 0 . 2 (Br-) 0 . 2 (Br-) 0.02 (I-)

1.1 0.2 0.2 0.03

(0-3.5) x 10-5 (3) (0-3.5) X 10-6 (4) (0.5-1.8) x (5) (1-10) x 10-8 (10)

1.0 f0.4 1.4 f 0 . 1 1 . 5 i: 0 . 1 7 . 2 It 0.8

* Number of trials in parentheses.

Ion pair; pH 2.

source of 12- for these experiments. It has previously mitted to stand more than about 30 min, so that some been established that the (Co(NH3)e3+,I-) ion pair is I- may be formed from the thermal hydrolysis,@or if efficiently photodecomposed into Co2+ and 12.26,34 M ) are introduced small amounts of I- ([I-] 5 The ammine complexes of cobalt(I1) equilibrate with into the solution before photolysis, then the Iz- transolvent in acidic aqueous solution rapidly enough that sient is always observed. The decay characteristics of they cannot be involved in a back reaction with 1 2 - or the 1 2 - transient formed from Co(NHa)J2+depend on the (estimated) ratio of [ I . ]t,o [Iz-]as expected,2 but I3-. I n this system we do observe that the pseudofirst-order rate constant for decay of 12-increases with do not depend on the [Co(NH3)J2+]remaining after flashing. When perchlorate solutions (from which I[Co(t~ans[l4]diene)~+]; however, following this reachas been excluded as rigorously as possible) of Cotion we observe a very slow increase in the absorptivity (“a)512+ are only partially (about 50% using a 50-J due to Ia-, This slower reaction is apparently the pulse) decomposed by flash photolysis, the substrate result of I- reduction of the (trans[l4]diene)cobalt(III) complexes formed in the initial step (8). The apparent absorbance at 390 nm changes only during the light pulse (pulse lifetimes were 10-20 psec); in our experisecond-order rate constant for this latter reaction is ments the substrate absorbance was observed to be conabout 35 A4-l sec-l under our reaction conditions. ~~ stant after 10-15 psec. Co(NH3)J2+. We have carefully r e - e ~ a m i n e dthe We found that solutions of Co(NHa)J?+ were too flash photolysis behavior of this complex to determine (1) if the radical i m p l i ~ a t e d ~ 6 ~in~the 6 ~ ~thermal 7 reducdifficult to handle to be convenient as a source of 12-. tion of Co(NH3),I2+can be identified as I . or 12- and (34) J. F. Endicott and M. Z. Hoffman, J . Phys. Chem., 70, 3389 (2) whether Co(NH3)d2+is a convenient source of IZ-. (1966). We find that if [Co(NH3)J](ClO& is dissolved in (35) S. A. Penkett and A. W. Adamson, J. Amer. Chem. Soc., 87, 0.1 M HC104, if all solutions are handled in the dark 2514 (1965). (36) A. Haim and H . Taube, ibid., 85, 495 (1963). and if solutions approximately lom6M in Co(NH3)J2+ (37) V. Balzani and V. Carassiti, “Photochemistry of Coordination are flashed in less than 15 min after dissolving the comCompounds,” Academic Press, New York, N. Y., 1970, p 205. plex, then no transient, absorbance is observed in the (38) F. Basolo and R. G . Pearson, “Mechanism of Inorganic Reacnear-ultraviolet spectral region. If solutions are pertions,” 2nd ed, Wiley, New York, N. Y ., 1967. The Journal of Physical Chemistry, Vol. 76,No. 16,19YB

PHOTOCHEMICAL BEHAVIOR OF COBALT COMPLEXES

Discussion

2227

Table 111: Reduction Potentials for Aqueous

Halogen and Dihalide Radicals A. Photoredox Behavior of Cobalt(III) Complexes Containing Macrocyclic Ligands. The estimated priEstimated reduction potentials, Va mary yield of about 0.02 for photoreduction of CoRadioal oouple X-C1 X=Br X = I (trans[l4]diene)CI2+ can be compared to yields of X. + e - = X 2.6 2.2 1.4b about 0.2 for C O ( N H ~ ) ; C ~ ~0.1 + , ~for ~ , ~trans-Co~ Xz- + e- -G= 2X-0 2.3 1.9 1.1 (en)&lz+ 25 and 0.03 for Co(tetraen)Cl+ l1 (tetraen is XZ+ e- S XZ0.6 0.3 0.3 tetraethylenepentamine). Although this variation in Using standard free energies of formation of X.(gas), and quantum yields cannot be easily rationalized at the X - ( a s ) as tabulated in ref 31. Standard free energies of solution present time, it is clear that all these systems must have of X. were approximated using solubility data of the nearest very efficient mechanisms for nonradiative relaxation noble gas (J. H. Hildebrand, J. M. Prausnitz, and R. L. Scott, “Regular and Related Solutions,” Van Nostrand-Reinhold, of their charge-transfer excited states; the efficiency of Princeton, N. J., 1970, p 204). b A value of 1.27 V was estimated deactivation varies only from SO to 98%. for this couple in ref 36. 0 For K = [XZ-]/[X.][X-] we have As we had expected, thermal back reactions are very used K cv lo6 M-1 in each case (ref 6b). For Brs-, .K has been important when macrocyclic complexes of cobalt (111) reported t o be greater than lo7 M-1 (B. Cercek, M. Ebert, C. W. Gilbert, and A. J. Swallow, “Pulse Radiolysis,” Academic are photoreduced. We have capitalized on this fact Press, New York, N. Y., 1965, p 83) and for Clz- greater than t o examine some features of dihalogen radical anion of equal t o lo2M-1 (ref 25). reactions with metal complexes. B. Dihalogen Radical Anion Reactions with Metal Complexes. The Xz- oxidations of Co(trans [14]erally more powerful oxidants than the X2 or X3- spediene)2+ all occur with second-order rate constants cies seems unimpeachable. which are very near to the diffusion-controlled limit. It is much more difficult to find direct evidence that Such variations in k8 as we have recorded in Table I1 X2- radical anions function as reducing agents. We appear to arise from differences in the ionic strength have found that the presence of Co(trans 114jdiene)rather than from differences in the reactivities of the does not change the Brz- decay rate. On the various XZ- species. When we commenced this study, other hand, the observation that the cobalt(I1) yield we had expected some variations in reactivity since the from irradiation of Co(NH3)J2+ is dependent on the Xz- radical anions vary in their oxidizing potentials incident light i n t e n ~ i t y ~ ~suggests * ~ ~ ~ 4 that 6 some such (Table 111) and the Co(trans[14]diene)(OH2)z3+~2+ XZ- reaction may occur in this ~ y s t e m ; ~thus e self-exchange rate is very slow (kco = 2 X 10v7 M-l H+ sec-’ a t 70°).39 Thus, in principle, values of kg might Co(NH3)J2+ hv +Co2+ 1. 5NH4+ (+la) have been used in connection with the R 4 a r c ~ scross ~~ I. I-ZIz(9) relation41 to determine the validity of potentials estimated in Table 111. More recently, however, it has been shown that the activation barriers observed in (39) D . P. Rillema, J. E”. Endicott, and N. A. P. Kane-Maguire J . Chem. SOC., Chem. Commun., 495 (1972). cobalt(II1)-(11) self-exchange reactions are not simply (40) For pertinent reviews see: (a) R. A. Marcus, Ann. Rev. Phys. related to the activation barriers observed in either reChem., 15, 155 (1965); (b) W. L. Reynolds and R. W. Lumry, ductions of cobalt(III)42 complexes or the oxidations of “Mechanisms of Electron Transfer,” Ronald Press, New York, N. Y., 1966. ~ o b a l t ( I 1 complexes. )~~ The value of k8 2 loe M-’ (41) In the present case, this would be ks = (kcak’Ksfs)’/2 where k’ sec-l observed for Xz- oxidations of Co(trans[l4]is a “self-exchange” rate constant for Iz/Ia-, Ks is the equilibrium constant for Ks and log f s = (log Ks)2/4 log (kook’/Z2) (for Z the diene)2+can be compared to the second-order rate concollision frequency). stant of approximately lo8 M-l sec-l observed for (42) D. P.Rillema, J. F. Endicott, and R. C. Patel, J. Amer. Chem. + both the CeIV (E” = 1.44 V)31and the R ~ ( b i p y ) ~ ~ SOC., 94,394 (1972); (b) J. F. Endicott, R. R. Schroeder, and D. R. (E” = 1.26 V)43 oxidations of this same substrate.17 Ferrier, J . Phys. Chem., submitted for publication; (c) D . P. Rillema and J. F. Endicott, rnn.org. Chem., in press. Allowing for the differences in reactant charge types (43) See D. A. Buckingham and A. M. Sargeson, “Chelating Agents there is less than two orders of magnitude difference in and Metal Chelates,” R. P. Dwyer and D. P. Mellor, Ed., Academic Press, New York, N. Y., 1964,Chapter 6,Bipy is bipyridine. these reactivities. (44) J. A. Kernohan, Ph.D. Dissertation, Boston University, 1969. One aspect of the potentials estimated in Table I11 Also note the similar inference recorded in the Results section, above. is clearly confirmed by this study. Despite the sim(45) Valentine (Advan. Photochem., 6, 123 (1968),footnote 64) has ilarity of potentials for the Co(trans[14]diene)attributed t o one of the present authors (J. F. E.) a statement that l6 and the Iz/I- 31 couples and despite the (OH2)23++2+ this intensity dependence is not real. At one time this author did verbally discuss the possibility of ambiguities resulting from the observation that excess I- reduces Co(trans [14]diene)techniques employed by Endicott and Hoffman.22 The effect in (OHZ)Z~+ we, ~find ~ that 1,- oxidizes Co(trans [14]question has been substantiated in the more careful work of Balzani and coworkers.87 diene)2+ (in excess iodide), demonstrating that Iz- is a (46) Haim and Taube86 have presented evidence that the enhanced significantly stronger oxidant than I,. The thermoreduction of Co(NHs)sIa+ is brought about by iodine atoms ( I . ) ; dynamic argument that X2- radical anions are gensee following discussion. 0

+

+

+ +

The Journal of Physical Chemistry, Vol. 78, No. 18, 1978

222s

SALLY D. MALONE AND JOHN F. ENDICOTT

+ C0(l\TH3)sIZ++ + CO'+ + 5KH4+ 212+ 1Iz- + I + IH+

12-

18-

41 3 -

(10) (11)

---t I2

(12)

21 + I2

(13)

We have therefore re-examined aspects of the photochemistry of Co(KH3)J2+ in order to determine whether evidence for reaction 10 can be found. Flash photolysis of Co(NH3)J2+ produces Iz- except when I- is rigorously excluded from the photolyte. When Izis observed, the amount is compatible with the thermal aquation of Co(NH3)J- 38 and the high stability of Iz(Kg = 8 X lo4 M-16b). Under flash conditions (ie., high I. and [Co(NH3)J2+]I 5 X M ) , we find no evidence for either (10) and (14). If these reactions 1.

+ CO(NH~)~IZ+ + CO'+ + ?in",+ Hf

---t 1 2

(14)

were diffusion controlled (Le., klo or k14 about 1O1O M-' sec-l), we should have observed a transient ( L I S > 10 psec) corresponding to the change in substrate absorbance (at 390 nm in the absence of I- or at 580 nm when I- was present), but we have not been able to detect such an absorbance change.47 We therefore conclude that reactions 10 and 14 are too slow to compete with (11)-(13) at such high radical concentrations. This is of course consistent with the previously observed intensity d e p e n d e n ~ e ~of~ ,the ~ ~ cobalt (11) yields since these continuous photolyses have einstein 1.-l indicated that even with I, N 5 X min-l (about 1/103 of our flash intensities) (10) and (14) are not competitive with (11)-(13). Although neither (10) nor (14) is observed directly, some relatively straightforward considerations suggest that observed intensity dependence of the cobalt(I1) yield can only arise from (10). In the continuous photolysis exM and 10-20 min periments [Co(NH3)J2+] = A after preparing such solutions [I-] N photostationary state treatment of [I.] and [L-] in the above mechanism leads to

-

Since ks, k12, and IC13 are all of the same magnitude (logto 1O1O M-l sec-1)2 and since [I-] lov5 M >> ([It-] [I.I), we have approximately (neglecting the contribution from [Iz-])@ [I 3 N +I,/(kg[I-] lc14[Co(NHa)512+]). Thus if 1. were the radical reattacking the substrate, the overall quantum yield, 4coz+= (4Ia k14[I*][Co(NH3),Iz-+]>/Ia should be independent of I,, and for the I-/Co(T\"3)J2+ reaction lc14 2 lo5M-l sec-l. From our flash photolysis studies we can only say that IC14 < 109 A I - l sec-l. However +coli. has been found to be I, dependent, so I. must not be the implicated radical. Our flash photolysis studies of CO(KH3)5I2+ show

+

+

+

The Journal of Physical Chemistry, Vol. 78, No. 16,1978

that: (1) no reduction of Co(NH3)J2+ (as monitored a t 390 and 580 nm) is observed (after the light8pulse) on the flash photolysis time scale, and (2) that 1 2 - is formed even when [I-] J c ~ ~ [ C O ( N H ~ then a photostationary state treatment gives

The proposed mechanism is compatible with the observed intensity dependence of 4coz+

= 41, + k0lI2-1I [Co(NH3)sIZ+] ,

if kio[Co("3)J2+I

N

+

(~ii[Iz-] hz4I~/.lcg[I-]]

when I, 'V einstein 1.-1 sec-l. Further substiM , [Co(KH3)J2+]N M , and tuting [I-] N k ~N z k7 N 1O'O M-' see-', and kll N 5 X logM-' sec-', M and klo N 2.5 X l o 4 M-I we find [Iz-1N 5 X sec-l. It is clear that the observed intensity dependence of 4 in this system can be related to a radical/ substrate reaction only if this reaction has a relatively small (compared t o kg,kll, klZ, or kI3)second-order rate const ant. Our very rough estimate of a reasonable value of klo may be compared to the observed value of 6.7 X lo3 M-' sec-' for the R u ( N H ~ ) ~(E" ~ += 0.1 V)49reduction of C O ( N H ~ ) J ~ +Since . ~ ~ the I&- self-exchange reaction must be slower than or equal to diffusion controlled (kex 5 1O1O AI-1 sec-l), the apparent similarities in the rates of the R ~ ( r \ l H ~ ) and 6 ~ +1 2 - re) ~ Iindicate ~+ that Iz- is a ductions of C O ( ~ ~ H ~does relatively good reducing agent; i.e., using the Marcus41 we estimate square root relation, klz = (k~lJc~2K1~)1'2, E" 5 0.45 V for the I&- ~ o ~ p l e . ~This ~ ~ ~ 7 ~ ~ limit is consistent with the estimate for this couple in Table 111. C . Mixed Dihalogen Radical Anions. The potentials estimated in Table I11 permit some inferences regarding the formation and stability of radical anions of the type XY-, where X @ Y. Such a radical anion (47) Note that our experimental conditions were adjusted so that the observed transmittance changed from about 0% T t o about 30% T during a single flash. This change in transmittance corresponds to about 50% decomposition of the Co(xHa)s12+ substrate. (48) The term k - g [ k - ] can make a significant contribution only if k12 is about the same magnitude as k ~ . Our experimental observations rule out this possibility. (49) T. J. Meyer and H. Taube, Inorg. Chem., 7 , 2369 (1968). (50) J. F. Endicott and H. Taube, J . Amer. Chem. Soc., 86, 1686 (1964). (51) Employed in this calculation are values of E o = 0.1 V48 and kexah = loa M-' sec-147.49 for the Ru(NHa)s3+/Ru(NHa)a2+couple. (52) G. Navon and D. Meyerstein, J . Phys. Chem., 74,4067 (1970).

BIPHOTONIC IONIZATION PROCESSES

2229

may be formed in either of two ways, (15) or (16). It follows then that the ratio of the corresponding formation constants, Kls and K16, and the first-order disso-

x. + Y - Z X Y x- + Y . -XYJ

(15) (16)

ciation rate constant, IC-15 and k-16, will depend on whether X. or Y . is the better oxidant; i.e.

As an example consider the case of IBr-. Clearly Br. is a far more powerful oxidant than I., and from the estimates in Table I11 we find [Bra ] [I-]/ [Br-1 [I.1 N 1.7 X Thus IBr- can only be formed from the reaction of Bra with I- and would be expected decay

into Br- and I. with a large first-order rate constant. For example, we find that when C O ( N H ~ ) ~isB flash ~~+ photolyzed in the presence of I- only the IC transient is observed. Finally, the mere observation of radical anions of the XY- type helps bracket the potential of the Y./Ycouples. Thus the observation of Schoneshofen and Henglei@ that both (1SCN)- and (BrSCN)- have appreciable lifetimes indicates that the NCS /NCScouple is pretty much central (E" N 1.8 V) between the I./I- and the Br./Br- couples. On the other hand, our observations that flash photolysis of Co(NH3)sNa2+ 420 in I- produces 12-, and a new transient (A, nm), presumably BrN3-, in 1.0 M Br- suggests that IS3.is only a little poorer oxidant than Br..

-

-

(53) (a) -M.Schoneshofen and A. Henplein, Ber. Bunsenges. Phys. Chem., 74, 393 (1970); (b) ibid., 73, 289 (1969).

Biphotonic Ionization Processes in Nonpolar Solutions of N,N,N',N'-Tetrameth yl-p-phenylenediamine by A. Alchalal, M. Tamir, and M. Ottolenghi* Department of Physical Chemistry, The Hebrew University, Jerusalem, Israel

(Received January 10, 1078)

Publication costs assisted by the U . S. National Bureau o j Standards

Photoconductivity experiments, using pulsed nitrogen-laser excitation a t 3371 1,are carried out in fluid . The effects of light inhydrocarbon solutions of T M P D (N,N,N',N'-tetramethyl-p-phenylenediamine) tensity and of added quenchers indicate that T M P D undergoes ionization in ci biphotonic process involving both excited singlet and triplet states as photoactive intermediates. T M P D solutions containing naphthalene and biphenyl as fluorescence quenchers are found to be photoconductive, exhibiting a second power dependence on the laser light intensity. The photocurrents are attributed to light absorption by the resulting fluorescent exciplexes. The details of this process are discussed in terms of the exciplex absorption spectrum.

Introduction Excited aromatic molecular systems in fluid solutions may undergo ionization via two alternative paths represented schematically by

interaction between an excited donor and a solute acceptor molecule (or vice versa) leading, in polar solvents, t o the solvated D + and A- radical ions.3

D* +D + 3. e,- (or R-)

(1) J. Jortner, M.Ottolenghi, and G. Stein, J . Amer. Chem. Soc., 85, 2712 (1963); J. Eloranta and H. binschitz, J . Chem. Phys., 38, 2214 (1963); L. I. Grossweiner and H. I. Joscheck, J . Amer. Chem. Soc., 88, 3261 (1966); H. S. Pilloff and A. C. Albrecht, Nature ( L o d o n ) , 212, 499 (1966). (2) C. Chachaty, D. Schoemaker, and R. Bensasson, Photochem. Photobiol., 12, 317 (1970); M. Ottolenghi, Chem. Phys. Lett., 12, 339 (1971). (3) H. Leonhardt and A. Weller, 2. Phys. Chem. (Frankfurt a m M a i n ) , 29, 277 (1961); Ber. Bunsenges. Phys. Chem., 67, 791 (1963); H. Knibbe, D. Rehm, and A. Weller, ibid., 72, 257 (1968); K. H. Grellmann, A. R. Watkins, and A. Weller, J . Luminescence, 1,2, 678 (1970).

(1)

D*+A+D++A(2) The first process involves electron transfer from an excited donor molecule (D*) to the solvent, followed by the formation of the D + radical cation and a solvated electron (e,-). I n some cases2 dissociative electron capture may occur, leading to the formation of solvent radical fragments (R-). Process 2 involves the

The Journal o j Physical Chemistry, Vol. 76, N o . 16, 1978