Anal. Chem. 2000, 72, 2828-2834
Photochemically Induced Energy-Transfer Effects on the Decay Times of Ruthenium Complexes in Polymers Paul Hartmann
AVL List GmbH, Hans-List-Platz 1, A-8020 Graz, Austria
The effects of photobleaching on absorption properties and decay times of ruthenium complexes adsorbed on silica gel or dissolved in polystyrene or PVC are reported. While complexes with bipyridyl ligands adsorbed on silica gel did not show any decay time decrease at all, complexes having phenanthroline ligands exhibit a significant photodegradation which is manifested not only by a decrease in luminescence intensity but also by decay time decrease and specific absorption spectral changes. The effects are shown to depend strongly on the oxygen and dye concentrations present at the bleaching process. The absorption bands of the photoproducts of ruthenium-phenanthroline complexes overlap with the emission spectra of the intact molecules. Fo1 rster resonance energy transfer from intact ruthenium complexes to their own photoproducts, generated by singlet oxygen attack of the phenanthroline bridges of the ligands, is suggested to be responsible for the observed decay time effects. Today, optical chemical sensors (optodes) are used for a variety of applications. Many of them are based on ruthenium complexes and serve as oxygen,1-4 pH,5,6 or temperature sensors.7 These dyes have favorable properties such as decay times of a few microseconds, absorption bands in a spectral range accessible by blue lightemitting diodes, high Stokes shift, and comparatively high photostability.8 In luminescence analysis, the decay time is frequently preferred to luminescence intensity as the information carrier: It is in a wide range independent of the concentration of the dye,9 and under certain conditions, photobleaching affects only emission intensities but not the decay time. (1) Wolfbeis, O. S.; Leiner, M. J. P.; Posch, H. E. Mikrochim. Acta 1986, III, 359-366. (2) Bacon, J. R.; Demas, J. N. Anal. Chem. 1987, 59, 2780-2785. (3) Preininger, C.; Klimant, I.; Wolfbeis, O. S. Anal. Chem. 1994, 66, 18411846. (4) Hartmann, P.; Leiner, M. J. P.; Lippitsch, M. E. Anal. Chem. 1995, 67, 88-93. (5) Murtaza, Z.; Chang, Q.; Rao, G.; Lin, H.; Lakowicz, J. R. Anal. Biochem. 1997, 247, 216-222. (6) Price, J. M.; Xu, W.; Demas, J. N.; DeGraff, B. A. Anal. Chem. 1998, 70, 265-270. (7) Demas, J. N.; DeGraff, B. A. Proc. SPIE-Int. Soc. Opt. Eng. 1992, 1796, 71-75. (8) Demas, J. N.; DeGraff, B. A. Anal. Chem. 1991, 63, 829A-837A. (9) Lakowicz, J. R. Principles of fluorescence spectroscopy; Plenum Press: New York and London, 1983.
2828 Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
However, this requires that the photoproducts are nonemissive and that there are no photoinduced changes of the dye’s microenvironment. The photochemistry of ruthenium complexes in a dilute solution is dominated by ligand loss and photoanation of one or more ligands by solvent molecules or counterions.10 The respective photoproducts have been indeed shown to be nonemissive at room temperature. This is also true for the reduced and oxidized species of many ruthenium complexes.11,12 Therefore, photoinduced changes of decay time parameters have not been observed in aqueous systems. Unfortunately, the conclusion that decay time is not affected by photobleaching is frequently generalized. Usually, polymer environments are chosen for optical sensor applications of luminescent dyes. In practice, rather thin layers and high dye concentrations are not uncommon to achieve the desired sensor properties (small volume, short response times, high luminescence signals). Some applications also require high excitation power or long-term illumination for continuous monitoring.13,14 Another aspect is that many of the dyes used for oxygen sensors also efficiently generate reactive singlet oxygen as a consequence of dynamic energy-transfer quenching of the indicator luminescence.15 Therefore, the photochemical stability of the sensors becomes increasingly important. Because of this, the photochemical stability of polymerimmobilized indicators has been addressed in previous work. Carraway et al.16 observed intensity photobleaching of a tris(4,7diphenyl-1,10-phenanthroline)-Ru(II) complex (RuPh2phen) mixed into filled silicones, but did not see any decay time changes. They concluded that singlet oxygen is probably not the primary cause of intensity loss of the sensor emission. Milosavljevic and Thomas17 found that high concentrations of tris(2,2′-bipyridyl)Ru(II) chloride (RuBpy, c > 5 mM) in cellulose (10) Durham, B.; Caspar, J. V.; Nagle, J. K.; Meyer, T. J. J. Am. Chem. Soc. 1982, 104, 4803-4810. (11) Kennelly, T.; Gafney, H. D.; Braun, M. J. Am. Chem. Soc. 1985, 107, 44314440. (12) Belser, P.; von Zelewsky, A.; Juris, A.; Barigelletti F.; Balzani, V. Gazz. Chim. Ital. 1983, 113, 731-735. (13) Klimant, I.; Holst, G. A.; Ku ¨ hl, M. Proc. SPIE-Int. Soc. Opt. Eng. 1995, 2508, 375-386. (14) Rosenzweig, Z.; Kopelman, R. Anal. Chem. 1995, 67, 2650-2654. (15) Demas, J. N.; Diemente, D.; Harris, E. W. J. Am. Chem. Soc. 1973, 95, 6864-6865. (16) Carraway, E. R.; Demas, J. N.; DeGraff, B. A.; Bacon, J. R. Anal. Chem. 1991, 63, 337-342. (17) Milosavljevic, B. H.; Thomas, J. K. J. Phys. Chem. 1983, 87, 616-621. 10.1021/ac9914723 CCC: $19.00
© 2000 American Chemical Society Published on Web 05/12/2000
led under high-intensity illumination to substantial concentration quenching and triplet-triplet annihilation effects (followed by disproportionation). This causes decay time changes upon irradiation, which increase with increasing concentration of the dye. Photoinduced disproportionation of RuBpy on porous Vycor glass (PVG) has been observed also by Kennelly et al.11 In this system, the RuBpy cation exchanges electrostatically onto the surface of the PVG. They reported on concentration effects leading to disproportionation reactions of dye molecules being located within the electron migration distance 1.3 nm < d 5 mM) in cellulose lead to substantial concentration quenching and triplet-triplet annihilation effects (followed by disproportionation). The latter effect may lead to stable reduction products (RuBpy3+), since a fraction of the emerged RuBpy33+ is reduced by the cellulose matrix before back reaction of the products can occur. This may lead to a decay time change via an additional reductive quenching pathway. Indeed, peaks corresponding to the reduced form of RuPh2phen have been observed in MALDI-TOF spectra of photobleached dyes.18 The increasing absorption bands near 540 nm of the absorption difference spectra may well be caused by appearance of the reduced form of the complex.25 But complexes having lost one or two ligands (this photochemical mechanism is effective also in solid solutions) are also expected to show up in this spectral region.10 The oxidized form absorbs only very weakly above 650 nm.26 The oxidized complex may also arise as a product of oxidative charge-transfer quenching by molecular oxygen, but Timpson et al.27 found that quenching of RuBpy by oxygen is best attributed to direct energy transfer yielding singlet oxygen, while the rate constant is only mediated by the presence of a low-lying chargetransfer state. However, redox reactions do not appear to be among the dominant mechanisms for the observed pronounced phase-shift changes: In fact, the investigated ruthenium complexes have almost identical electrochemical potentials28 and should therefore all have a similar redox photochemistry. But in practice they show a highly different photochemical behavior; the RuBpy/silica gel system does not show any decay time change at all. The dye and oxygen dependence of the photoeffects suggests rather that self-sensitized singlet oxygen attack of the dye is dominant. Indeed, there are reports, for example, on the selfsensitized photodecomposition of rubrene in polystyrene by singlet oxygen.29 This molecular system is in some respect similar to the oxygen sensors used in the present study. At the present stage, it is most likely that a Fo¨rster resonance energy-transfer mechanism is responsible for the decay time changes of phenanthroline-based ruthenium complexes immobilized in polymers at relatively high concentrations: Dynamic oxygen quenching (eq 4) yields reactive singlet oxygen, which attacks nearby ruthenium complexes having phenanthroline ligands (eq 5). Singlet oxygen is expected to react predominantly with the phenanthroline 5,6-bridge16 due to the comparatively low π-electron density. This process is even more pronounced for 4,7-diphenyl-1,10-phenanthroline ligands, while
RuBpy molecules that are lacking the 5,6-bridge are not affected. The resulting photoproducts have not been identified unambiguously. Results from MALDI-TOF measurements on bleached samples18,30 possibly point to a dioxygen addition reaction of the ligands leading via a dioxetane intermediate to a stable aldehyde (Figure 9). In any case, the photoproducts are nonemissive and characterized by pronounced absorption bands around 600 nm. These bands overlap with the emission spectra of intact ruthenium complexes (Figure 4). At sufficiently high concentrations, a Fo¨rster resonance energy-transfer mechanism (eq 6) between the intact molecules (acting as donors), and the photoproducts (acting as acceptors) effectively reduces the measured decay times by increasing the nonradiative decay rate and finally converting the excited-state energy into heat.
RuL3*2+ + 3O2 f RuL32+ + 1O2
(4)
RuL32+ + 1O2 f RuL2(L′O2)
(5)
RuL3*2+ + RuL2(L′O2) f RuL32+ + RuL2(L′O2)* (6) where
L ) Phen, respectively, Ph2phen; L′ ) modified ligands To confirm this hypothesis, we have applied several singlet oxygen scavengers to our sensor systems18 and were able to show that, for example, the known singlet oxygen scavenger DABCO reduces the photoinduced decrease of the decay time of RuPh2phen/PS sensors substantially, while the decrease of the luminescence intensity is inhibited only to a small extent. This confirms also that the intensity decrease is probably dominated by ligand loss, in agreement with the findings of Carraway et al.16 The observed dependence of the photobleaching rate on the polymer may also be related to singlet oxygen: The environment of singlet oxygen (e.g., the polymer) is able to provide an acceptor of the excited singlet-state energy by its vibrational modes. This depends strongly on the molecular composition of the environment (C-H and O-H stretch vibrations are the most effective sinks), which is reflected by the intrinsic decay time of singlet oxygen in various polymers. The lifetime of the oxygen singlet state in PS is rather low compared to other polymers (τ ) 19 (25) Creutz, C.; Sutin, N. J. Am. Chem. Soc. 1976, 98, 6384-6385. (26) Balzani, V.; Bolletta, F.; Gandolfi, M. T.; Maestri, M. Top. Curr. Chem., 1978, 75, 1-64. (27) Timpson, C. J.; Carter, C. C.; Olmsted, J., III J. Phys. Chem. 1989, 93, 41164120. (28) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85-277. (29) Ogilby, P. R.; Kristiansen, M.; Martire, D. O.; Scurlock, R. D.; Taylor, V. L.; Clough, R. L.; Adv. Chem. Ser. 1996, No. 249, 113-126. (30) Roth, T. Abstr. Europt(r)ode IV Conf., Mu ¨ nster 1998, 225-226.
Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
2833
µs).29 Therefore, PS acts as a relatively efficient singlet oxygen scavenger, which may limit singlet oxygen attack of the ruthenium complex. But within the lifetime of 19 µs, the activated oxygen molecules may diffuse as much as d ) 43 nm in PS (given a diffusion constant of D ) 7.7 × 10-7cm2/s31) and still attack dye molecules nearby (at c ) 5 mM the average distance between dye molecules is ∼7 nm). The stretched exponential decay law (eq 3) has been alternatively suggested to describe the decay data of ruthenium complexes in various polymer layers.22 In view of the photochemical mechanism proposed in the present work, this decay law appears in a straightforward way as the consequence of ongoing degradation of the dye molecules. This may also be responsible for the extraordinary high deviations from single-exponential profiles (i.e., high values of a0) reported by Draxler et al. for similar sensor systems.22 A possible effect of the microheterogeneous environment is comparatively small, at least in PVC, as evidenced by the near-single-exponential luminescence decay of freshly prepared RuPh2phen/PVC sensors, while in PS, the RuPh2phen emission of freshly prepared samples shows a higher degree of nonexponentiality. Special care must also be taken of the preparation of
the RuPh2phen perchlorate dye, since decomposition of the molecules may already take place during synthesis owing to the use of hypochloric acid.
(31) Wang, B.; Ogilby, P. R. Can. J. Chem. 1995, 73, 1831-1840.
AC9914723
2834
Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
CONCLUSIONS The presumed benefits of decay time techniques do not apply in every case and have to be critically investigated in view of the necessary boundary conditions, as, for example, the complex microenvironment of the dyes in polymers and the frequently used high dye concentrations of many sensor systems. However, the reported results point to successful strategies of optimization of the sensor performance with respect to improved long-term stability and photostability. ACKNOWLEDGMENT The author acknowledges helpful discussions with M. Leiner, F. Andreae (Graz), T. Roth (Zurich), and I. Klimant (Regensburg), as well as technical support by P. Kohlbacher and A. Boila-Go¨ckel (Graz). Received for review December 22, 1999. Accepted March 28, 2000.
