Chromium(II)-polypyridyl complexes: formation, spectra, and electron

Nick Serpone, Mary A. Jamieson, Salvatore S. Emmi, Pier G. Fuochi, Quinto G. Mulazzani, and Morton Z. Hoffman. J. Am. Chem. Soc. , 1981, 103 (5), pp 1...
0 downloads 0 Views 988KB Size
J. Am. Chem. SOC.1981, 103, 1091-1098

1091

Chromium( 11)-Polypyridyl Complexes: Formation, Spectra, and Electron-Transfer Kinetics Nick Serpone,*la.bMary A. Jamieson,laSalvatore S. Emmi,*lCPier G. Fuochi,lc Quinto G. MuIazzani,lCand Morton Z. Hoffman*ld Contribution from the Department of Chemistry, Concordia University, Montreal, Quebec, Canada H 3 G lM8, the Laboratorio di Fotochimica e Radiazioni d’Alta Energia del C.N.R., 40126 Bologna, Italy, and the Department of Chemistry, Boston University, Boston, Massachusetts 02215. Received April 24, 1980

Abstract: Flash photolysis techniques have been employed to obtain kinetic data for electron transfer between the couples Cr(NN),Z+-Fe3+ (as) at -22 OC in Nz-purged 1 M HCI, aqueous solution for N N = bpy, 4,4‘-Phzbpy, phen, 5-Me(phen), 5-Cl(phen), 5-Br(phen), 5-Ph(phen), and 4,7-Phzphen (under conditions of total quenching of the l u m i n w n t 2Estate of Cr(NN)$+ by Fez+ (as)). Also for N N = 4,4‘-Me2bpy, 5,6-Me2phen, and 4,7-Mezphen, pulse radiolysis techniques were used under pseudo-first-order conditions in Fe’+(aq) in Ar-purged solutions (-25 OC, ionic strength 0.15 M, pH -1) with the Cr(I1) transients formed by selective reduction of the Cr(NN),,+ complexes by the a-2-propanol radical, (CH3)zc0H. The rates are nearly diffusion-controlled 108-10’o M-I s-l, Application of the Marcus relationships for electron-transfer processes yields k, c 108-109 M-I s-l for both the self-exchange rate in the Cr(NN)33+-Cr(NN)32+ couple and for the couple (’E/’T1) Cr(NN),’+-Cr(NN)?+. The similar k, value has been interpreted in terms of identical inner-sphere barriers for electron exchange; *AGin( = AGi,’. Near-UV and visible absorption spectra of the polypyridylchromium(I1) transients, obtained by both flash photolysis (Nz-purged aqueous solutions) and pulse radiolysis (Ar-purged methanolic solutions), are reported and interpreted in terms of simple ligand field theory.

Introduction Although Cr(bpy)?+ (bpy = 2J’-bipyridine) has been used for many years as a reducing agent in inorganic electron-transfer reactions, interest in the properties of polypyridyl complexes of chromium(II), Cr(NN)32+: has increased dramatically recently due to the observation that Cr(NN)?+ can be generated photochemically from the reductive quenching of the 2E excited state of Cr(NN),,+., In particular, recognizing the potential utility of Cr(NN),,+ complexes in solar energy conversion and storage scheme^,^,^ Cr(NN)32+ species could serve a s potential charge carriers in photogalvanic cells6 and as mediators with sufficient potential to reduce H20. In fact, it has been reported that Cr( b p ~ ) , ~reduces + H 2 0 to H2 in acidic media7,*and disproportionates to Cr(bpy),+ and Cr(bpy),’+ in alkaline solution^.^ We have been engaged in a detailed study of the photochemistry and photophysics of the 2E excited state of Cr(NN),,+ complexes which is metal-centered and l ~ m i n e s c e n t . ~ . ’As ~ part of this systematic approach toward the understanding of the competitive modes of energy partitioning in excited states, we have focused (1) (a) Concordia University. (b) Visiting Professor, Spring Semester 1979, Boston University. (c) FRAE, CNR Bologna. (d) Boston University. (2) The following ligand abbreviations have been used: bpy = 2,2’-bipyridine; 4,4’-Me2bpy = 4,4’-dimethyl-2,2’-bipyridine; 4,4’-Ph2bpy = 4,4‘diphenyl-2,2’-bipyridine; phen = 1,lO-phenanthroline; 5-Cl(phen) = 5chloro-1,IO-phenanthroline; 5-Br(phen) = 5-bromo-1,IO-phenanthroline;5Me(phen) = 5-methyl-1,lO-phenanthroline;5-Ph(phen) = 5-phenyl-1,lOphenanthroline; 5,6-Me2phen = 5,6-dimethyl-l,IO-phenanthroline; 4,7Melphen = 4,7-dimethyl-l,lO-phenanthroline;and 4,7-Ph2phen = 4,7-diphenyl- 1,IO-phenanthroline. (3) N. Serpone, M. A. Jamieson, M. S.Henry, M. Z. Hoffman, F. Bolletta, and M. Maestri, J. Am. Chem. Soc., 101,2907 (1979), and references therein. (4) (a) N. Serpone and M. Z. Hoffman, ACS/CSJ Chemical Congress, Honolulu, HI, April 1979, Abstract INOR-446. (b) N. Serpone and M. Z. Hoffman, Chemical Institute of Canada Conference, Vancouver, B. C., June 1979, Abstract IN-25. (5) N. Serpone and M. Z. Hoffman, to be submitted for publication. (6) M. W. Hersey, T.J. Vandermoot, and C. H. Langford, Inorg. Chim. Acta, 29, L233 (1978). (7) R. Murray and G. M. Waind, Proceedings of the 7th International Conference on Coordination Chemistry, Stockholm, Sweden, 1962, p 309. (8) D. S.Miller and G. McLendon, ACS National Meeting, Washington, D. C., September 1979, INOR- 1 19. (9) S. Herzog and K. C. Reimer, Chem. Ber., 92, 872 (1959). (10) R. Sriram, M. 2.Hoffman, M. A. Jamieson, and N. Serpone, J. Am. Chem. Soc., 102, 1754 (1980). 0002-7863/8 1/1503-1091$01.25/0

this study on the formation of C T ( N N ) , ~ +by the reductive quenching of (2E) Cr(NN)33+with Fe*+(aq) and by the direct with radiation-generated free radicals. reduction of CT(NN)~,+ We report here the absorption spectra of Cr(NN)32+transients and the kinetics of electron back-transfer between Cr(NN)J2+ and Fe3+(aq)studied by use of the techniques of flash photolysis and pulse radiolysis.

Experimental Section

Complexes. The following complexes were available from an earlier study: IC~(~PY),I(C~O~)~.~/ZHZ~, [C~(~,~’-P~Z~PY),I(C~O~),.~HZ [Cr(4,4’-Me2bpy)3](C104)3.Hz0, [Cr(phen),](C104),.2Hz0, [Cr(S-CI[Cr(4,7(phen))3](C104)3-2Hz0,[Cr(4,7-Mezphen)3](C104)3.2Hz0, Ph2phen),] (C104)3.4Hz0,and [Cr(3,4,7,8-Me4phen),] (C104)3.2.5Hz0. The following ligands were obtained from G.F. Smith and Co. and used as received: 5-methyl-1,lO-phenanthroline(5-Me(phen)), 5,6-dimethyl-1,lO-phenanthroline (5,6-Me2phen), S-phenyl-l,lOphenanthroline (5-Ph(phen)), and 5-bromo- 1, IO-phenanthroline (5-Br(phen)). The preparation of new Cr(NN),’+ complexes was carried out according to procedures established earlier.’ Elemental analyses were performed by Galbraith Laboratories, Inc., Knoxville, TN. [C~(~-B~(~~~II))~](CIO~),.~.~H~O Absorption spectrum in 0.1 M HCI, X (log t) 213 nm (4.97), 240 (4.85), 275 (4.75), 367 (sh, 3.55), 436 (sh, 2.88), 470 (sh, 2.58). Anal. Calcd C, 36.31; H, 2.37; N, 7.06; CI, 8.93; Cr, 4.37; Br, 20.13. Found: C, 36.24; H, 2.24; N, 7.03; C1, 9.13; Cr, 4.55; Br, 19.98. [Cr(!LMe(phen))3](C104)3-2H20 Absorption spectrum in 1 M HCI, X (log e) 212 nm (5.06), 225 (sh, 5.01), 275 (4.92), 310 (sh, 4.36), 370 (sh, 3.61). 440 (sh, 2.75), 470 (sh, 2.47). Anal. Calcd: C, 48.34; H, 3.54; N, 8.67; CI, 10.97; Cr, 5.36. Found: C, 48.68; H, 3.52; N, 8.65; C1, 10.89; Cr, 5.16. [Cr(5-Ph(phen)),](CI04),.2.5H20 Absorption spectrum in 1 M HC1 (with -5% v/v MeOH), A (log e) 227 nm (5.02), 243 (4.97), 279 (4.87), 370 (sh, 3.76), 470 (sh, 2.71). Anal. Calcd: C, 55.71; H, 3.55; N, 7.22; C1,9.13;Cr,4.46. Found: C,55.52;H,3.55;N,7.36;C1,9.31;Cr,4.19. [Cr(5,aMe~)3](~4)3,1.~z~ Absorption spectrum in 1 M HCI, X (log t) 213 nm (4.94), 243 (4.85), 279 (4.77), 320 (sh, 4.21), 380 (sh, 3.53), 470 (sh, 2.58). Anal. Calcd: C, 50.34; H, 3.92; N, 8.31; C1, 10.52; Cr, 5.19. Found: C, 50.20; H, 3.56; N, 8.24; C1, 10.80; Cr, 4.99. Materials. Fe(NH4)2(S04)2.6H20(containing 0.005% F e 9 , NaC104, THF, anhydrous CH30H, 95% C2H50H, HCI (0.1 ppm Fe), HC104 (60%, 0.2 ppm Fe), H2S04, Ar gas (IGT99.998%), and Fe(NH,)(SO4)2 were of reagent quality. Monodistilled water was used or was further purified by distillation from acidic CrzO:- and alkaline Mn04and fractionation in an all-silica apparatus. For the pulse radiolysis experiments, CH,OH (Merck or C. Erba) was purified by the method

0 1981 American Chemical Society

1092 J . Am. Chem. SOC.,Vol. 103, No. 5, 1981

Serpone et al.

of Baxendale and Wardman;" 2-propanol (Baker Instra Analyzed) was used as received. Nitrous oxide (IGT 99.98%) was passed through NaOH pellets just prior to use. Apparatus. Absorption spectra of the Cr(II1) complexes were recorded with an Aminco-Bowman DW-2 UV-vis spectrophotometer. Roomtemperature luminescence spectra were taken with a Perkin-Elmer MPF-2A spectrofluorimeter equipped with a R-446 photomultiplier tube and high-intensity accessory. Flash photolysis experiments were performed by using an apparatus previously described.I2 Emission lifetimes were determined by using a 1 kW N2 laser (4-11s pulses at 337 nm); the luminescence decay curves were photographed from a Tektronix 7633 oscilloscope and were exponential in all cases (correlation coefficient >99.9%). Pulse radiolyses with optical absorption detection were performed by using the apparatus at the FRAE-CNR Laboratory in Bologna (Italy)." Radiation doses of 0.5-2 krad were used in the determination of transient absorption spectra and 0.1-0.5 krad was employed in the study of electron-transfer kinetics. Procedures. Unless otherwise specified, photochemical experiments were carried out in 1 M HCI aqueous solution at 22-23 "C;radiation chemical experiments were performed at -25 OC. Solutions were deoxygenated by purging with N2 or Ar and saturating with N 2 0 (2.5 X M, H20, 25 O C , 1 atm)14 when required. Solutions of limited solubility in water were dissolved in a small amount of CH30H (- 1%) followed by addition of 1 M HCI or were examined in methanolic solution. Steady-state luminescence and quenching experiments were carried out with right-angle illumination on 10-5-10" M solutions of the newly prepared complexes. The excitation wavelength was chosen such that the absorbance of the solution in 1-cm cells was < O X In the luminescence quenching experiments with Fe2'(aq) the solutions were air-equilibrated; stock solutions of Fe2'(aq) were prepared immediately prior to use. Where appropriate, the ratio I 0 / I was corrected for absorption of light by Fe2+(aq). In the flash photolysis experiments, the xenon flash ( l / e decay time = 30 ps) was filtered through acetone (A > 330 nm) contained in the outer jacket of a quartz cell (22-cm path length). The analyzing light was filtered through a monochromator (400-700-nm range). The N2-purged solutions were replaced after every second flash; no changes in the solution absorption spectra were detected as a result of the flash under the experimental conditions used. For the determination of the Cr(NN),2t transient spectra, the appropriate solution of Fe(NH4)2(S04)2-6H20for >99% quenching of the (2E) Cr(NN)33t was freshly prepared and added to the Cr(NN)," solution, and the mixture was immediately deoxygenated to prevent any air oxidation of Fezt(aq). In the pulse radiolysis experiments, the samples were protected from the photolytic effects of the analyzing light by using an appropriate UV cutoff filter (A > 400 nm). In every case, the thermal stability of the Cr(NN),)' solution was verified over the period of the experiments. The rates of electron transfer between Cr(NN)pzt and Fest(aq), where NN = 5,6-Me2phen, 4,7-Me2phen, and 4,4'-Me2bpy, were determined under pseudo-first-order conditions with Fe-"(aq) in excess; when [Cr(NN)?'] = 1-3 pM (radiation doses of 0.16-0.45 krad), [Fe3'(aq)] = 35-70 pM. The rates were independent of the dose. The disappearance of the absorbance due to C T ( N N ) ~was ~ + monitored at 420 nm with the ionic strength held constant at 0.15 M. The spectra of Cr(NN)32+were obtained from the reduction of Cr(NN),)' by the CY-(CH,)~COHradical at pH -1; [Cr(NN),3t] = 40-60 gM containing 0.05 M H2S04and 0.52 M 2-propanol. Generation of Reducing Radicals. The radiolysis of aqueous solutions results in the decomposition of the solvent and the generation of the primary radicals according to the following stoichiometry: H 2 0 e,, (2.7), OH (2.7), H (0.55) (where the numbers in parentheses represent the number of species formed per 100 eV of energy absorbed by the solution (G value)). In acidic solution, eaq-is efficiently scavenged by Ht, yielding H atoms (3.65). In the presence of 2-propano1, H and OH react to form a- and @-(CH3)2COHradicals; the a radical is a rather good reducing agent (Elj20(NHE) = -1.2 V).15 The concentration of (CH,),cOH generated in the pulse was obtained from the expression [(CH,),COH] = DfGp where D = dose in krad,f= 1,036 X 10" for

--

(11) J. H. Baxendale and P. Wardman, J . Chem. SOC.,Faraday Trans.

1 , 69, 584 (1973).

(12) A. F. Vando, E. R. Kantrowitz, M. Z. Hoffman, E. Papaconstantinou, and J. F. Endicott, J. Am. Chem. Soc., 94, 6655 (1972). (13) A. Hutton, G.Roffi, and A. Martelli, paper presented at the 5th Congress of Radiation Research, Seattle, Washington, July 1974. Quad. dell' Area Ric. dell'Emilia-Romagna, 5, 67 (1974). (14) W. F. Linke, "Solubilities. Inorganic and Metal-Organic Compounds", Vol. I, 4th ed.,American Chemical Society, Washington, D.C. 1958, Vol. 2, 1965. (15) J. F. Endicott in "Concepts of Inorganic Photochemistry", A. W. Adamson and P. D. Fleischauer, Eds., Wiley-Interscience, 1975, p 88.

Table I. Photophysical Data of New Cr(NN),3+Complexes in 1 M HCl Solutions Z Ex 10-3 *T, x io-, NN To,a PS cm-' cm-'

a

5-Br(phen) 5-Ph(phen) 5-Me(phen) 5,6-Me,phen [CI(NN),~+]*

68

13.74 14.3 13.70 14.3 59 13.70 14.3 40 13.66 14.3 M; air-equilibrated solutions; 22 "C. 55

4-

3

0

W

O ' 2 4 0 '

'

370

'

'

450

I

530 '

'

'

610

'

690

"

A,nm

Figure 1. Absorption spectra of Cr(bpy):' (ref 29) (dashed line, 0 (this work)) and of the Cr(4,4'-Me2bpy)3Zt ( 0 )transient obtained by reduction of the corresponding Cr(II1) complex (53 1M) with (CH3)2COH at pH 1 in Ar-purged aqueous solutions containing 0.26 M 2-propanol and 0.05 M H2SO4 (dose/pulse = 1.30 krad; G(Cr") = 5.5).

-

H 2 0 (and methanolic) solutions irradiated by an electron beam, G = G value for the radical (5.5), and p = density of the solution. In CH30H solutions, G = 2 for em,; and p = 0.79 g/cm3.

Results Table I summarizes the photophysical data for the newly synthesized Cr(NN)?+: lifetime of 2E state in air-saturated solution a t 22 O C and the energies of the luminescent ZEand *TI states. Comparison of these values with those published for the other complexes shows the close similarity in the behavior and energetics of the lower doublet excited states. The rate constants for the quenching of the ZEluminescence, kq, by Fez+ of the new complexes is reported in Table I1 together with those values for the other c o m p l e ~ e s .The ~ values of k, were obtained from linear Stern-Volmer plots of Zo/Z = 1 + ksv[Fe2+(aq)l = 1 + kqrair [Fe2+(aq)], where rairis the lifetime of ZEin air-saturated 1 M HCl aqueous solution. Also shown in Table I1 are the kinetics of the reduction of Cr(NN)33+by ea, and by ( Y - ( C H ~ ) ~ C Oand H the wavelength at which formation of Cr(NN):+ was monitored. The 3,4,7,8-Me4phen complex was too insoluble in aqueous and methanolic solutions to afford the spectrum of the Cr(I1) species and the kinetics of its oxidation by Fe3+(aq). In Table I1 are also summarized the rate constants for the reaction of Cr(NN)32+and Fe3+(aq). In the flash photolysis experiments, where Cr(NN)32+ is generated by the reductive quenching of (zE)Cr(NN)33+by Fez+(aq), the decay of Cr(NN)32+is via second-order kinetics. From knowledge of the absorption spectra of the Cr(NN)3z+ species (see below), the second-order rate constant was obtained. For the cases where N N = 4,4'-Mezbpy, 5,6-Mezphen, and 4,7Me2phen, the rate constants were also obtained by pulse radiolysis through the selective reduction of Cr(II1) in the presence of excess Fe3+(aq) followed by pseudo-first-order decay of Cr(NN)32+. Figures 1 through 10 illustrate the absorption spectra ( 3 W 7 5 0 nm) of the Cr(NN)32+complexes obtained from either the reorHthe quenching duction of Cr(NN)?+ by e,,,- or ( Y - ( C H ~ ) ~ C O of (zE) Cr(NN)33+by Fe2+(aq). Values of e were obtained from the pulse radiolysis experiments where the concentration of Cr-

J. Am. Chem. SOC.,Vol. 103, No. 5, 1981 1093

Chromium(II)-Polypyridyl Complexes.

gq0

a

7Vtn

0.4

R -0.3

-

- 0.2

x 10

u

lt

I

0.3

A00

I

I l l

1

290

370

I

I

I

I

530

450

I

I

610

i

I 4 1 0 690

h.nm

Figure 2. Absorption spectra of the Cr(4,4'-Phzbpy)3z+ transient: (0)

by flash photolysis methods in N,-purged 1 M HC1 aqueous solutions containing 5% v/v MeUH (see experimental section); ( 0 )by pulse radiolysis methods in which the Cr(II1) species (45 pM) were reduced by e&- at natural pH in Ar-purged methanolic solutions (dose/pulse = 1.47 krad; G(Cr") = 1.5). I

I

X,nm

Figure 5. Absorption spectra of Cr(S-Br(~hen))~~+ ions: (0)by flash

photolysis in N2-purged aqueous 1 M HCI solutions (see Experimental Section); ( 0 )by pulse radiolysis by reduction of Cr(II1) species (39 pM) with (CH3)&OH at pH -1 in Ar-purged aqueous solutions containing 0.52 M 2-propanol and 0.05M H 3 0 , (dose/pulse = 0.54 krad; G(CrI1) = 5.1).

--

X,nm

Figure 3. Absorption spectra of Cr(phen)32+: (0)by flash photolysis in Nz-purged 1 M HCI aqueous solutions (see Experimental Section); ( 0 ) by pulse radiolysis methods by reduction of the Cr(II1) species (52 pM)

with (CH3),COH at pH -2 in Ar-purged aqueous solutions containing 0.26 M 2-propanol and 0.01 M HC10, (dose/pulse = 1.27 krad; G(CrlI) = 4.6). I

'

'

'

I'

h

41

'

'

'

'

'

'

'

' I

O L '

I

290

1

1

'

370

'

'

450

1

I

530

'

1

610

'

1

690

1

1

1

I

370

'

450

1

'

530

1

'

610

1

'

690

10

X .nm Figure 6. Absorption spectra of Cr(S-Me(~hen))~~+: (0)by flash photolysis techniques in Nz-purged aqueous 1 M HCI solutions (see Experimental Section); ( 0 )by pulse radiolysis techniques by reduction of Cr(II1) species (52 pM) with (CH3)#0H at pH -1 in Ar-purged aqueous solutions containing 0.26 M 2-propanol and 0.05 M HISO, (dose/pulse = 0.57 had; G(Cr") = 5.2).

4t

40.7

0'

1

290

I

1

X,nm

Figure 4. Absorption spectra of the Cr(S-Cl(phen),'+ ion: (0)by flash photolysis in N,-purged 1 M HCI aqueous solutions (see Experimental Section); ( 0 )by pulse radiolysis by reduction of Cr(II1) species (50 pM) with (CH3)#OH at pH 1 in Ar-purged aqueous solutions containing 0.9 M 2-propanol and 0.1 M HCIO4 (dose/pulse = 1.8 krad; G(CrI1) =

-

4.8).

(NN)?' was established from the quantitative scavenging of the

reducing radical. In the case of the 5-Ph(phen) complex, we report

290

370

450

530 X,nm

610

690

Figure 7. Absorption spectra of the Cr(S-Ph(~hen))~~+ species: (0)by

flash photolysis in N,-purged 1 M HCI aqueous solutions containing 1% v/v MeOH (see Experimental Section); ( 0 )by pulse radiolysis methods by reduction of the Cr(II1) species (49 pM) with ed; in Ar-purged methanolic solutions at natural pH (dose/pulse = 1.20 krad; G(Cr") =

2) *

1094 J . Am. Chem. SOC.,Vol. 103, No. 5, 1981 I

I

I

I

l

l

I

I

I

I

Serpone et al.

I

I

I

- 0.5

I/

I

+ AOC

b

- 0.3 - 0.1

I

I

I

I

I

I

I

I

I

I

I

4

370

290

450

530

590

610

X,nm

Figure 8. Absorption spectra of the Cr(5,6-Me2phen)32+ ion: (0)by flash photolysis in N2-purged aqueous 1 M HC1 solutions (see Experimental Section); ( 0 )by pulse radiolysis methods by reduction of Cr(II1) species (53 pM) with (CH3)2c0Hat pH -1 in Ar-purged aqueous solutions containing 0.26 M 2-propanol and 0.05 M H2S04(dose/pulse = 1.37 krad; G(Crl*) = 4.9).

11

10

P

7

5

0

.-_ - - - -

...

I/?;,

,

/

w

,

,

'I

r

,

,

,

,

,

,I

REOUCT I V E O l l i N C H I NG

x10

I\tt I

290

011

'

6

10

8

12

14 LOG

I

I

3 70

450

I

I

530

I

I

610

I

I

690

I

X,nm

Figure 9. Absorption spectrum of Cr(4,7-Me2phen)t+ species obtained by pulse radiolysis techniques by reduction of the corresponding Cr(II1) species (54 pM) with e& in Ar-purged methanolic solutions at natural pH (dose/pulse = 1.67 krad; G(Cr") = 1.8).

only AOD vs. A; the spectrum was taken 1 ps after the pulse corresponding to -90% formation of Cr(NN):+. After this time, however, an additional reaction was observed, perhaps involving radical attack on the ligand, which yields products absorbing in the same spectral region. Thus, for the 5-Ph(phen) complex, the calculation of e values carries some uncertainty. In Table I11 we have summarized the spectral characteristics of the Cr(NN)32+ complexes.

Discussion Excitation of ground-state (4A2)Cr(NN)33+into the spin-allowed quartet manifold forms (4T2)Cr(NN)$+ which, upon intersystem crossing, produces (2E) Cr(NN)33+. This long-lived, metal-centered, luminescent state can be quenched by Fe2+(aq) in competition with its radiative, nonradiative, reactive, and ground-state quenching modes of decay. At low [substrate] (10-5-10" M), ground-state quenching is negligible; in 1 M HC1, stable the complexes are thermally'6*17and (16) M. Maestri, F. Bolletta, N. Serpone, L. Moggi, and V. Balzani, Inorg. Chem., 15, 2048 (1976). (17) M. A. Jamieson, N. Serpone, and M. Maestri, Inorg. Chem., 17,2432 (1978).

6

16

13

20

22

K12

Figure 11. Log k vs. log KI2plot ( 0 )for the quenching of *Cr(NN)l+ excited states by heq2+ in air-equilibrated, 1 M HCI aqueous solutions; (0) log kI2 vs. log K12plot for the reduction of Fee:+ by Cr(NN)32+ transients (see text). The dashed line is the calculated curve for k22 = lo8 M-I s-l (see ref 25) and the solid line, calculated curve for k2, = loJ M-' s-l (see Discussion). The numbers indicate the following NN: (1) bpy, (2) 4,4'-Ph2bpy, (3) 4,4'-Me2bpy, (4) 5-Cl(phen), (5) S-Br(phen), (6) 5-Ph(phen), (7) 5-Me(phen), (8) phen, (9) 5,6-Me2phen, (10) 4,7Ph2phen, (1 1) 4,7-Me2phen.

with respect to the formation of the aquated product. Quenching of 2E by Fe2+(aq). In our previous study3 of the excited-state behavior of Cr(NN)?+ complexes, we concluded that the quenching of (2E) Cr(NN)$+ by Fe2+(aq) is via electrontransfer reaction 1; excluded were oxidative quenching by Fe2+(aq) (2E) Cr(NN)33+ Fe2+(aq) Cr(NN)32+ Fe3+(aq) (1) or electronic energy tran~fer.~JIThe observation of the formation of the spectra of C r ( N N ) t + and the dependence of k on the value of *Eo[(2E) Cr(NN)33+/Cr(NN)32+1]affords evilence for reductive electron transfer. The Marcus relationships for electron transfer22 are usually f ~ r m u l a t e das~ ~k12= (k11k22K1h2)1/2 and logf,, = (log K12),/4 log (kllk22/Z2)where kll and kz2 refer to the rate constants of

+

-

+

(18) R.Sham, M. S. Henry, and M. Z. Hoffman, Inorg. Chem., 18,1727 (1979). (19) M. Maestri, F. Bolletta, L. Moggi, V. Balzani, M. S.Henry, and M. Z. Hoffman, J. Am. Chem. SOC.,100, 2694 (1978). (20) M. A. Jamieson, N. Serpone, M. Maestri, F. Bolletta, V. Balzani, M. S. Henry, and M. Z. Hoffman, to be submitted for publication. (21) B. Brunschwig and N. Sutin, J. Am. Chem. Soc., 100,7568 (1978). (22) R. A. Marcus, J . Chem. Phys., 43, 2654, 679 (1965). (23) C. Creutz, Inorg. Chem., 17, 1046 (1978).

J . Am. Chem. SOC.,Vol. 103, No. 5, 1981 1095

Chromium(II)-PolypyridylComplexes.

Table 11. Kinetics of Electron-Transfer Processes in Chromium-Polypyridyl Complexes in 1 M HCl Aqueous Solutions (-22-24 "C) 10-7kq: M-,S-1 'E + Fez+(aq)

complexes

*E", v (,E)

Cr(III)/Cr(II)

10+k, M-, scl reducing agent + Cr(II1)

E", v

10-9k M-, s-1 cr(IIj + Fe"

C~(III)/C~(II)

1.44 + 0.04 -0.26b 1.6 f 0.2 (-0.28r -0.45 10.6 + 0.6 0.70 + 0.02' -0.17b (CH,),cOH; 2.8 (400 nm) 0.61 f 0.08 Cr(S-Cl(phen)), 3 t 12b (4.QC 1.53b -0.15g (CH,),COH; 3.8 (420 nm) 1.4 + 0.4 , Cr(5-Br(phen)), ,+ 8.4 1.55 Cr(5-Ph(phen)), ,+ 4.5 1.49 esolv-; 50-80 (1.9 + 0.6)' -0.21k -0.308 (CH,),cOH; 3.1 (400 nm) 2.0 ? 0.7 Cr(5-Me(phen)),,+ 3.4d (1.0)C 1.40 -0.28b 1.3 f 0.3 Cr(phen), ,+ 3.2b 1.42b (CH,),COH; 4.1 (480 nm) -0.2gk 2.7 + 1.0 2.6 1.40 (CH,),cOH; 3.6 (420 nm) Cr(5,6-Me2phen),'+ 0.48 + 0.03' (1.41b esolv-; 82 (575 nm) 1.4 f 0.4 (- 0.26r Cr(4,7-Ph2phen), ,+ 2.5 Cr(4,7-Me2phen),,+ 0.60b (0.089)c 1.24 (7)' -0.45 esolv-; 81 (575 nm) 0.71 + 0.09' (-0.5 7)e (l.ll)e h Cr(3,4,7,8-Me4phen),,+ ( 18)'~ 0.092b Phosphorescence quenching; air-saturated solutions; [Cr(NN),''] = M. Reference 3. In 1 M H2S04.21 In 1 M H,SO,, k, = 1 x 107 M - 1 S - 1 .2 1 e Calculated by interpolation from log k,, obtained from the log k, vs. log K,,plot of Figure 11. See also footnote 63 in ref 3. fWavelength at which kinetics were monitored. Reference 21. This complex was too insoluble in solvents used for pulse radiolysis. pH -1; ionic strength 0.15; in Ar-purged solutions containing 0.05 M H,SO, and 0.52 M 2-propanoL This value may be uncertain due to uncertainties in the E values of C1(5-Ph(phen)),~+(see text). Vs NHE; calculated taking E" of Cr(phen),3t/Cr(phen),2+ as -0.28 (ref 21) and using AEf values for the appropriate Cr(III)/Cr(II) couples (see Table IV of M. C. Hughes and D. J. Macero, Znorg. CRem., 13, 2739 (1974)). Calculated by interpolation from the plot of log k v s log K,, and included the complexes for whichE"'s were calculated as noted in footnote e. The concentration of Fe3+in Fe(NH4),(S0,),~6H,O was too high to permit the observation of the Cr(I1) transient and therefore to measure the rate of the electron-transfer process (see text). 3.7b (1.6)' 2.0b 0.22b (0.082)c

C ~ ( ~ P Y,+) ,

Cr(4,4'-Ph2bpy), '+ Cr(4,4'-Me2bpy), 3t

1.44b (1.39r 1.25

the Fe2+(aq)-Fe3+(aq) and (2E) Cr(NN)33+-Cr(NN)32+selfexchange reactions, respectively, k12 is the rate constant for the cross-reaction (kp of reaction l ) , and Z is the collision number (generally taken as 10" M-' S-I). The value of the rate constant (k22)of reaction 2 can be evaluated from k l l (4.0 M-I s-1)24 and (2E) Cr(NN)33+ Cr(NN)32+ Cr(NN)?+ (2E) Cr(NN),,+ (2) the plot in Figure 11 (open circles). Using the relationship log k12= 0.50 log kllk22+ 0.50(1 + a)log Ki2, where a = (log K12)/4 log (kllk22/Z2), we obtain an apparent kZ2 = lo5 M-' s-1.25 Recently, Sutin and Creutz26have demonstrated that although reactions of poly(pyridine)metal complexes with Fe2+(aq) give a free-energy plot having the proper slope, the intercept a t Ki2 = 1 is less than expected by several orders of magnitude. Endicott and Ferraudi2' have also estimated k22 to be 3 X lo5 M-' s-' . In a related study, Brunschwig and Sutin2' estimated k22as 1 X lo8 M-' s-l in 1 M H2S04a t 25 O C . Experimentally, however, the intercept of the line from which the exchange rate was evaluated was 1.7 from which k, = 600 M-' s-' is calculated. The 6 orders of magnitude discrepancy between the experimental and theoretically estimated values was explained in terms of the nonadiabaticity of the cross-reaction (1) while the self-exchange reactions are probably adiabatic. The Marcus relationship was modified2I to include a term for the probability of electron transfer where in the activated complex;28k!2.= p12(kllk22K12fi2/p1~22)i/2, p are the respective probabilities of electron transfer in the activated complexes. Clearly, this expression reduces to the usual Marcus equation when pI2= pii = p22 = 1, that is, when the self-exchange reactions and the cross-reaction are adiabatic. From above, pil = pZz= 1 but p12 450 nm. To the extent that Cr(NN)?+ are low-spin d4 complexes (3T1 the spin-allowed band ground state in 0,mi~rosymmetry),~' maxima may be estimated by second-order perturbation theory using the equations3*given by Konig and H e r ~ o g .In~ the ~ case of the analogous Cr(NN)33+complexes, we previously indicated (31) P. M. Lutz, G. J. Long, and W. A. Baker, Jr., Inorg. Chem., 8,2529 (1969). (32) With the assumption that C 4B, the energy maxima for the transitions are as follows: VI 'EJa), A - 2 8 64B2/A - 5.338; 'T 'TS(a), A + 52B2/A - 4.708; 'TI 'T2(a), A + 2B 52B2/A - 6.258; 'TI A,, A - 8 64B2/A; 'TI 'A2, A 3 8 - 80B2 A; 'TI 'T (b), A 148 12B2/A + 6.258; 'TI 'E(b), A 8 46B /A 5.338; 'TI )Tl(b), A 168 - 36B2/A 4.708.

+

+

+

--

+

++ +

+

I +

--

- -+

that the value for the ligand field parameter A is nearly the same for the three bpy complexes (-23.6 X lo3 ~ m - ' ) . ~Using the empirical r e l a t i o n ~ h i p A ~ ~=Jigand gi, where gi, = 17.4 X lo3 cm-', we obtain&,, = 1.36, in good agreement with the value of 1.33 given by J ~ r g e n s e n .The ~ ~ gvalue for low-spin Cr(I1) is 12.4 X lo3 cm-' 34 which, with thefligand factor, predicts A = 16.9 X lo3 cm-' for the three Crl*-bpy complexes. Tentative assignments for these bpy complexes are presented in Table I11 and are based on a comparison between experimental band maxima and calculated energies by using the above rela(33) C. K. Jorgensen, "Oxidation Numbers and Oxidation States", Springer, New York, 1969. (34) C. K. Jorgensen, "Absorption Spectra and Chemical Bonding in Complexes", Pergamon Press, Oxford, 1962.

J. Am. Chem. SOC.,Vol. 103, No. 5, 1981 1097

Chromium(II)-Polypyridyl Complexes.

ti on ship^^^ with the assumption that the Racah parameter B is constant in this series. The assignments for Cr(bpy)?+ are based on those given by Kdnig and Herzog except that we assign the 3T1 3T1(b) and 3T1 'T2(b) transitions to the bands a t 360 and 339 nm, respectively. In the Cr(4,4'-Me2bpy)?+ spectrum, these two transitions occur at -360 nm and are not resolved due to the very intense intraligand transition a t 305 nm. It is inter3Tz(a) esting, however, that the transitions 3T1 'Tl(a) and are resolved in this spectrum but are not in the bpy analogue. In the Cr(4,4'-Ph2bpy)? spectrum, the very intense bands at 270 and 330 nm are ascribed to intraligand r-r* transitions on the basis of intensity and by comparison with the spectrum of the protonated form of the free ligand which has bands a t 260 and 3 12 11113.~~ The much richer spectrum of the diphenyl complex is probably due to either vibrational components3 or to transitions predominantly charge transfer in nature. The low-energy transition 3T1 3E(a) is expected to occur at X > 700 nm in the Me2and Ph2bpyCr" spectra. (b) phen Spectra. Unlike the spectra of the bpy complexes, those of the Cr"phen complexes reveal relatively weak, barely discernible bands a t -520-650 nm. It is likely that transitions in this region have little or no charge-transfer character. At X < 520 nm and > 650 nm, intense bands are observed and, therefore, must be, to a large extent, charge transfer in nature. Earlier we reported A = 23.4 X lo3 cm-' for the 4,7-Me2phen, 3,4,7,8-Me4phen, and phen complexes of Cr(III), 22.2 X lo3 cm-' for the 5-Cl(phen) complex and 21.6 X lo3 cm-' for the 4,7Ph2phen c o m p l e ~ .From ~ the absorption spectra of the new Cr(II1) complexes reported here, A = 22 X lo3 cm-I for the 5-Br(phen)and 5-Me(phen)CrI1* complexes and -21.3 X lo3 cm-' for the 5-Ph(phen) and 5,6-Mezphen analogues. Taking g(Cr(II1)) as 17.4 X l o 3 cm-' 33 and the above A values, fii and = 1.34 for 4,7-Me2phen, 3,4,7,8-Me4phen, and phen, in excebent agreement with the value of 1.34 reported for phen by J0rgense1-1,~~ 1.27 for the 5-C1-, 5-Br-, and 5-Me-substituted phen, and 1.23 for 4,7Ph2phen, 5-Ph(phen), and 5,6-Mezphen. With the value of g for Cr(II),34 the values of A for the Cr(NN)?+ species are 16.6 X lo3 cm-' for N N = 4,7-Mezphen, 3,4,7,8-Me4phen, and phen, 15.7 X lo3 cm-' for N N = 5-Cl(phen), 5-Br(phen), and 5-Me(phen), and 15.3 X lo3 cm-' for N N = 4.7-Ph,phen, 5-Ph(phen), and 5,6-Mezphen. To estimate the energy maxima on the basis of ligand field consideration^,^^ we have made the assumption that the value of the Racah parameter B (-600 cm-') for these phen complexes is nearly the same as that of the bpy complexes. This ~ +Niis not unreasonable since the B values for N i ( b ~ y ) ~and hen)^^' are nearly identical (898 and 908 cm-', r e s p e ~ t i v e l y ) . ~ ~ The tentative assignments in Table I11 have been based on such estimates. The analysis indicates that the 3T1 3Tl(b) and 3T2(b) transitions should be degenerate in all of the spectra of the phen complexes. This is not generally observed (cf. Table 111). W e note that transitions to the 3E(a) excited states are not seen in the spectrum of the phen complex, which we attribute to the band 3E(a) maxima lying a t A > 700 nm. The 3T1 3Tl(a) and transitions are expected a t X > 700 nm in the spectrum of the 5-substituted phen complexes; the extra bands a t -400-560 nm have not been assigned. In general, substituents a t the 5-phen position red-shift the corresponding transitions in the parent phen complex by -20-40 nm; the 5-Br and 5-Ph substituents effect a larger red shift. Methyl and phenyl substituents a t the 5,6- and 4,7-positions on the phen ring system also red-shift the transitions of Cr(phen)?'. Phenyl substituents at the 4,7-position and 5,6-dimethyl substituents are particularly effective, to the extent that 3T1 3T2(a), 3Tl(a), and 3E(a) occur a t X > 700 nm. Additional unassigned bands are also evident in the spectra of these disubstituted phen complexes. It is likely that these bands also possess d - d character and the rich spectra probably originate from

transitions between states of the trigonal symmetry, D3, of the complexes. Reaction of Cr(NN)32+with Fe3+(aq). Reaction 5 represents

(35) Sadtler Index of UV Spectra, Sadtter Research Laboratories, Inc., 1967. (36) J. Jasephsen and C. E. Schaffer, Acta Chem. Scund., Ser. A, A31 813 (1977).

(37) G. J. Ferraudi and J. F. Endicott, Inorg. Chim.Acta, 37, 219 (1979). (38) N. A. P. Kane-Maguire, J. Conway, and C. H. Langford, J . Chem. Soc., Chem. Commun., 801 (1974). (39) a and j3 represent electron spins in the three tZgorbitals.

-

-

-

-

-

-

-

-

-

-

Cr(NN)32+

+ Fe3+(aq)

-

Cr(NN)33+

+ Fe2+(aq)

(5)

the electron back-transfer process between the products of the excited-state quenching reaction (1). Insofar as reaction 1 or its analogue involving other reductive quenchers is of potential importance in solar energy schemes? knowledge of the rate constants of reaction 5, kkt, as a function of ligand structure and kinetic parameters is necessary. As seen in Table 11, the values of kbt span the range 108-1010 M-'s-l with the slowest rate shown by the 5-Cl(phen) complex and the fastest by the 3,4,7,8-Me4phen complex; the latter rate was estimated from the plot in Figure 1 1. Most of the rate constants were obtained by the flash photolysis technique from the second-order decay of Cr(NN)?+ according to reaction 5 ; Cr(NN)32+and Fe3+(aq) are generated in equimolar amounts in reaction 1. For the 4,7-Me2phen and 3,4,7,8-Me4phen complexes, the low values (- lo6 M-' s-' ) of t h e rate constant of reaction 1 (kq) require rather large concentrations of Fe2+(aq) in order to quench completely (>99%) (2E) Cr(NN)?+. Unfortunately, the concentration of Fe3+(aq) impurity in the Fe(NH4)2(S04)2used is sufficiently high such that reaction 5 shows pseudo-first-order kinetics, too rapid to be followed within the time resolution of the flash photolysis apparatus (-30 p s ) . Instead, the pulse radiolytic technique was employed with reactions 2 and 3 forming Cr(NN)32+in the presence of a n excess of Fe3+(aq) relative to Cr(NN)32+; the kinetic concentration of Fe3+(aq) was always less than that of Cr(NN)33+. The values of kbct for the 4,4'-Me2bpy, 4,7-Mezphen, and 5,6-Mezphen complexes were obtained a t p H 1 and an ionic strength of 0.15 M. The plot of Figure 11, based on the data of k a t 1 M ionic strength and the values of EO(Cr(NN)J'+/Cr(NN)?+) from Table 11, provides the estimates of kkt for the 4,7-Me2phen and 3,4,7,8-Me4phen complexes in 1 M HCl. Comparison of these data and those taken at 1 M ionic strength for the same complexes reveals an order of magnitude difference which we attribute to the differences in solution medium. The data for kkt can also be treated by the Marcus relationships. The plot of log k k , vs. log K I 2(closed circles, Figure 11) also yields an apparent self-exchange rate constant for reaction 6 of of 10' M-' s-l with the actual value 108-109 M-' s-I (see

-

+

-

-

(4A2) Cr(NN)33+ Cr(NN)32+ Cr(NN)32+

+ (4A2) Cr(NN)33+ (6)

above discussion). Ferraudi and Endicott3' have reported k6 to be -4 X lo9 M-' s-l (for N N = bpy) in 90% C H 3 0 H at 25 O C and a t an ionic strength of 0.1 M. Irrespective of this difference, which may reflect the different solution medium and the errors intrinsic in the technique, it must be noted that, for very similar conditions of solution medium, temperature, and ionic strength, the rate of the excited-state self-exchange reaction (2) is identical with the corresponding rate of the ground-state self-exchange reaction (6). To a first approximation, the work terms and solvent reorganization barriers for reactions 2 and 6 are expected to be the same for the following reasons: (1) the complexes involved in both reactions have the same charges; (2) (2E) Cr(NN)33+and (4A2) C T ( N N ) ~are ~ + isogeometric inasmuch as the Stokes shift between the 2E 4A2emission3*and absorption29bands is zero and their orbital population is identical (t22), although different in spin [(a)(a)(p)vs. (a)(c~)(a)].~~ Consequently, if both the groundstate and excited-state reactions are adiabatic, they can only differ in their inner-sphere barriers.21 Differences in these barriers for the ground-state (AGi,') and excited-state (*AGi:) exchange reactions have been approximated2' by eq 7 for metal-centered excited states, wherefis a breathing force constant, (I is the Cr-N

-

1098

J . Am. Chem. SOC.1981, 103, 1098-1102

-

A(AGi,,*) = *AGi,( - AGi,,*

fz + f 3 (7) bond distance, and the subscripts 2 and 3 denote the charges on the complexes. Inasmuch as both 2E and 4A2states have the same t$3 configuration, the first term in eq 7 is expected to be small.@ Five possible cases have been describedz1 for the second term of equation 7. (i) E, = 0 (Le., a3* - a3 = 0), so that A(AGi:) also equals 0 and the inner-sphere barriers for the ground-state and excited-state reactions are the same. This suggests that the exchange rate for these two reactions should be nearly identical, as is observed. Also, (ii) If a3* > a3 and az > ' / 2 ( ~ 3 * a3),then *AGin* < AGi,,*. (iii) If a3* > a3 but a2 < I / 2 ( ~ 3 * + a3),then *AGh* > AGh*. (iv) If a3* < a3but az > '/Z(a3* a), then *AGh* > AGi:. (v) If a3* < a3 but a2 < l/z(a3* + a3),then *AGi,) < AGh*. The cases ii-v, which are dependent on whether the excited state has Cr-N bond lengths longer than or shorter than those in the ground-state species, are inconsequential in the present system. However, it is clear that if electron exchange involved addition of electrons to antibonding eg* orbitals, significant changes in the bond lengths and breathing force constants would take p l a ~ e . ~ ,Interestingly, ~' ZEdoes not appear to be the state involved in the reductive quenching reaction (8); a higher energy state has been suggested but not identified.37

+

+

+

-

+

* C r ( b ~ y ) ~ ~R+O H C r ( b p ~ ) ~ ~ROH' + (8) Involvement by the 4Tzstate in excited-state self-exchange reactions can be ruled out due to its very short lifetime (