J. Phys. Chem. B 2007, 111, 12075-12080
12075
Triplet-Polaron Quenching in Conjugated Polymers D. Hertel* and K. Meerholz Institute of Physical Chemistry, UniVersity of Cologne, Luxemburgerstrasse 116, 50939 Cologne, Germany ReceiVed: July 16, 2007; In Final Form: August 9, 2007
We studied the triplet-polaron quenching in a platinum(II) porphyrin- (PtOEP-) doped polyspirobifluorene (PSF-TAD) copolymer. The copolymer contains a hole-transporting phenylenediamine unit (TAD) as a comonomer. Triplet-polaron quenching was probed by the change in PtOEP phosphorescence lifetime under an applied voltage in a unipolar device. The charge-induced reduction of the optically excited lifetime of PtOEP is one-third for the highest applied bias. The charge density can be obtained from current-voltage characteristics in the space-charge-limited (SCL) regime. The obtained hole mobility under SCL conditions is (7 ( 2) × 10-5 cm2/(V s). This result is in accord with recent mobility measurements of the time-of-flight mobility in our polymer. The triplet-polaron recombination constant was evaluated to be (4 ( 1) × 10-13 cm3/s, implying a triplet-polaron interaction radius of 2 × 10-10 m. The results show that triplet-polaron annihilation cannot be neglected in device models for phosphorescent light-emitting diodes.
1. Introduction The triplet state of conjugated polymers has recently attracted great attention1-3 because of its important role in the photophysics of these materials. In conventional conjugated polymers, intersystem crossing from the singlet (S1) to the triplet state (T1) or from T1 to S0 (phosphorescence) is spin-forbidden; hence, at room temperature, only fluorescence emission is observed. The pioneering work of Forrest and co-workers4-6 has shown that triplet excitons can be used to enhance the efficiency of OLEDs by harvesting through energy transfer6 or by using molecules with heavy-metal atoms exhibiting spin-orbit coupling as phosphorescent emitters.4,5 Meanwhile, the efficiency of green-emitting phosphorescent devices is close to the theoretical limit for quantum efficiency.5,7,8 The concept has been shown to work equally well for solution-processed polymer OLEDs,9 although polymer devices are more difficult to optimize than vapor-deposited OLEDs. This is mainly due to the complicated fabrication of multilayer devices. Despite the success of the phosphorescent-emitter concept in polymer OLEDs, the picture of photophysical processes involving the triplet state is far from complete. A triplet emitter doped into a polymer matrix is populated by energy transfer after photoexcitation or by direct recombination of charges under electrical excitation.4,6,10 Because of their long lifetimes, large amounts of triplets accumulate in the device. The high triplet concentration leads to triplet-triplet annihilation (TTA), commonly observed in conjugated polymers11,12 and a major loss mechanism in phosphorescent OLEDs.13 A model describing the fate of triplets in polymer OLEDs is lacking. This is, at least in part, due to the fact that, for a quantitative description, the rate constants for various processes are needed. Few attempts have been made to determine TTA rate constants in polymers. In polyfluorene derivatives, TTA has been studied in detail by phosphorescence11,14,15 and optically detected resonance techniques.16-18 The latter usually require multiparameter fits, as most of the input variables are not known a priori. To date, * To whom correspondence should be addressed. E-mail: dirk.hertel@ uni-koeln.de.
less attention has been paid to triplet-polaron annihilation (TPA). A polaron denotes a charge carrier or radical ion. Because the charge-carrier mobility in polymer materials is low, a significant accumulation of charges in a device close to its capacitor charge occurs. Hence, one can expect an influence of TPA on the photophysics of triplet excitons in devices. Very recently,19 it has been proposed that TPA plays a major role in the organic magnetoresistance effect.20,21 From a more fundamental point of view, the mechanistic details of triplet-polaron interactions are largely unknown. In this work, we report a study of the triplet-polaron interactions in a conjugated polyspirobifluorene polymer doped with the phosphorescent emitter platinum porphyrin (PtOEP). The phosphorescence lifetime of optical excited PtOEP can be used as a sensitive probe of TPA. The change of PtOEP lifetime with voltage (current) allows us to quantify triplet-polaron interactions. A similar method was successfully applied to molecular crystals decades ago.22,23 To vary the amount of charges in our device quantitatively, we made use of spacecharge-limited currents (SCLCs). SCLCs can be achieved by insertion of cross-linkable hole-transport layers. Thus, we were able to obtain quantitative information about triplet-polaron quenching in our devices. In addition to enabling the estimation of charge density, SCLC allows the mobility to be determined and compared to recent time-of-flight measurements.24 It was found that TPA can be analyzed in terms of a diffusioncontrolled reaction, yielding an interaction radius of 2 × 10-10 m. 2. Results and Discussion Photophysical Properties of PSF-TAD/PtOEP. Before proceeding to the study of the triplet-polaron interactions in devices, we briefly discuss the photophysical properties of PSFTAD to show that it fulfils all requirements for our investigations. The emission spectra of the materials were measured with gated detection techniques using an intensified CCD. A dye laser was used to excite the samples at either 400 nm (PSF-TAD) or 536 nm (PtOEP). Figure 1 displays the absorption and delayed emission spectra of PSF-TAD. The phosphorescence spectrum
10.1021/jp075556o CCC: $37.00 © 2007 American Chemical Society Published on Web 10/05/2007
12076 J. Phys. Chem. B, Vol. 111, No. 42, 2007
Hertel and Meerholz
Figure 1. Absorption spectrum (black line, 298 K) and delayed emission spectrum (line with data points) of PSF-TAD. Spectra were recorded with a time delay of 10 ms after excitation at 3.49 eV at 80 K. The high-energy part of the spectrum (blue) is delayed fluorescence; the red part corresponds to phosphorescence. For comparison, the phosphorescence of PtOEP (dotted line, 298 K) detected 1 µs after excitation at 2.31 eV is shown as well. The inset shows the structure of the polymer used, PSF-TAD. The ratio of the two monomers is 1:1.
of PtOEP is included for comparison. The absorption of the PSF-TAD film is broad and structureless, typical for disordered solids, with a maximum at 3.24 eV (383 nm). Fluorescence (80 K) with the S0 r S1 (0-0) transition at 2.7 eV (459 nm) is characterized by a vibronic splitting of 150 meV. This vibronic splitting is a convolution of several vibronic modes, as is often observed in conjugated polymers.25 At room temperature, the fluorescence is slightly blue-shifted to 2.8 eV (442 nm). The delayed fluorescence (DF) spectrum coincides with the prompt fluorescence as expected. The source of DF is TTA, as in other polyfluorene-type polymers,11,12,15 and will be discussed elsewhere. At long times after optical excitation, the phosphorescence of PSF-TAD can be observed with the S0 r T1 (0-0) transition at 2.14 eV (579 nm). Apparently, there is no emission from the TAD unit, which would be expected at 2.38 eV (520 nm) according to the assumption that the energy of the T1 state is similar to that in triphenylenediamine (TPD), which is chemically closely related to TAD. By comparison of the T1 energy of PSF-TAD to that of PtOEP (Figure 1), it is apparent that PtOEP constitutes a trap for triplet excitons in the PSFTAD matrix. Therefore, no energy transfer is expected from the PtOEP dopant to the PSF-TAD matrix, which is a prerequisite for our investigation. The dopant can be excited directly at 536 nm without excitation of the matrix, as there is no absorption of PSF-TAD. Current-Induced Phosphorescence Lifetime Quenching. To investigate triplet-polaron annihilation (TPA), we studied the time-resolved emission of PtOEP in PSF-TAD under an applied bias. The unipolar device used consisted of a glass substrate covered with an indium tin oxide (ITO) anode; a layer of PEDOT/PF, a cross-linkable hole-transport layer (AUPD; see Experimental Section); and PSF-TAD doped with PtOEP. As a cathode, a gold/silver layer was used. PEDOT/PF was used instead of the conventional PEDOT/PSS material because of
the higher work function. PF is a fluorinated polymeric counterion.26,27 To synchronize the applied voltage with the laser excitation (10 Hz), a pulsed rather than a continuous voltage was utilized, thereby reducing the thermal stress imposed on the sample. The voltage pulse duration was about 600 µs. The decay of the PtOEP phosphorescence at 650 nm in the unipolar device is shown in Figure 2. To fit the decay, a singleexponential function was used. This gives a very good agreement.28 The lifetime of PtOEP at zero applied voltage is 69 µs. In thin films on quartz substrates (without metal electrodes), the lifetime is 100 µs in accord with literature data.29,30 To exclude temperature effects as a source for the shortening of the lifetime, we studied the temperature dependence of PtOEP phosphorescence in the temperature range from 250 to 320 K and found no change in lifetime within experimental error ((1%). The reduction in lifetime by one-third is related to the quenching of triplet excitons by the metal surface.31 Because of the long lifetime of the triplet exciton, significant diffusion can be expected even at dopant concentrations of 1%. However, if a voltage is applied to the unipolar device, a substantial reduction of the PtOEP lifetime can be detected (Figure 2). At 16 V (E ) 2.7 × 105 V/cm), the phosphorescence lifetime was reduced to 52 µs. For simplicity, single-exponential decays were used to fit the data. The electric field can be excluded as a source of the decrease in phosphorescence lifetime as borne out by the inset of Figure 2. In this control device, the cathode is silver, and an HTL (OTPD) with a higher oxidation potential33 is used to block hole injection. It is obvious that the PtOEP lifetime of the control device remains unchanged (70 µs, inset Figure 2) even at twice the electric field strength compared to that of the unipolar device in Figure 2. Kinetics of Triplet States. To analyze our findings quantitatively, let us briefly recall the kinetics of triplet excitons in molecular solids.32 If triplets are generated upon direct excitation
Triplet-Polaron Quenching in Conjugated Polymers
J. Phys. Chem. B, Vol. 111, No. 42, 2007 12077
Figure 2. Plot of the logarithm of the phosphorescence intensity of PtOEP (at 650 nm) doped into PSF-TAD in a unipolar device as a function of time. The decay curves correspond to applied voltages of (b) 0, (4) 5, (1) 10, and (0) 16 V. The device (ITO/PEDOT/PF/AUPD/ PSF-TAD/1% PtOEP/Au/Ag) has a thickness of 590 nm (E ) 2.7 × 105 V/cm for 16 V). The solid lines are single-exponential fits corresponding to lifetimes of (b) 69 and (0) 52 µs. The inset shows the logarithm of the phosphorescence intensity of PtOEP vs time for a reference device. The decay was measured with (O) 0 and (s) -20 V applied bias. The lifetime in both samples was 70 µs. The device was 285 nm thick (E ) 7 × 105 V/cm for 20 V).
or via ISC from an optically excited singlet state, they can decay in a monomolecular or bimolecular fashion. The rate equation for the triplet concentration, [T], can be expressed as1
d[T] ) -k0[T] - γTTA[T]2 - γTPA[T]nh dt
(1)
where k0 is the sum of the radiative (kr) and nonradiative (knr) decays and, hence, the inverse phosphorescence lifetime (k0 ) knr + kr ) 1/τphos). γTTA is the bimolecular triplet-triplet annihilation constant. The last term in eq 1 includes losses due to bimolecular annihilation of triplets with polarons. γTPA is the rate constant for this process, and nh is the concentration of charges (in our case, holes) in the device. Under the experimental conditions used here, i.e., low excitation density, the TTA term can be neglected. This simplifies eq 1 to
d[T] ) -k0[T] - γTPA[T]nh ) -k[T] dt
(2)
The overall decay rate of phosphorescence, k, in the presence of charges is then given by
k ) k0 + γTPAnh
(3)
or, expressing the decay rate as difference of rates (quenching) normalized to the intrinsic decay rate k0
k - k0 γTPAnh ) k0 k0
(4)
Thus, the measurement of the phosphorescence lifetime allows for extraction of γTPA if the charge density is known. Assuming
Figure 3. Normalized change of the phosphorescence decay rate of PtOEP (1% in PSF-TAD device) as a function of voltage applied to the unipolar device (525-nm-thick) at 315 K. The line corresponds to a linear fit. The inset shows the quenching efficiency of several devices as a function of organic layer thickness on double logarithmic scales. The fit corresponds to a quadratic dependence according to eq 5.
that the charge density, nh, is given by the space-charge density in the device, nh in eq 4 can be replaced
k - k0 γTPA[(3rV)/(2ed2)] ) k0 k0
(5)
Here, V is the applied voltage, d is the sample thickness, is the permittivity of free space, and r is the dielectric constant. According to eq 5, the increase in quenching, i.e., decrease in lifetime, should vary linearly with voltage if it is due to the space-charge density. The corresponding plot is shown for a 590-nm-thick PtOEP/PSF-TAD device in Figure 3. The dependence of the lifetime quenching on thickness (Figure 3 inset) is another test for the validity of the application of eq 5. The agreement of both measurements with eq 5 is striking. Space-Charge-Limited Current. To verify that the current in our device is space-charge-limited, we turn now to an analysis of the current voltage (IV) characteristics under pulsed electrical excitation measured simultaneously with the phosphorescence lifetime. The devices were constructed to facilitate ohmic injection of holes into the PSF-TAD material by the use of the AUPD. The HOMO of AUPD is at -5.3 eV,33,34 and the HOMOs of PSF-TAD are at -5.39 and -5.63 eV24 for the TAD and PSF units, respectively. PtOEP has no influence on charge transport because its HOMO is lower (-5.6 eV) and does not constitute a hole trap. For a unipolar device, we expect that the dependence of hole-current density j on applied voltage V can be described by35
j)
9r µV2 8ed3
(6)
where d is the sample thickness, µ is the hole mobility, is the permittivity of free space, and r is the dielectric constant. For a number of devices, the current-voltage (IV) characteristics taken under pulsed conditions are shown in Figure 4. The quadratic dependence of the current density on the voltage expected for SCLC behavior is apparent. The inset shows the thickness dependence of the current density, again in accord with SCLC. One should note that these IV curves were measured
12078 J. Phys. Chem. B, Vol. 111, No. 42, 2007
Hertel and Meerholz simultaneously achieve a relatively high rate constant. So far, our measurements yield a TPA constant under conditions relevant for devices. The magnitude of the TPA constant shows that it is in the same range as TTA. Thus, TPA has to be included in the photophysical picture for device models. The quenching of phosphorescence by charges accounts for about one-third of the loss in the quantum efficiency of OLEDs based on T1 emitters having similar lifetimes. We can go one step further in the analysis of our data. The theory of bimoleculardiffusion-controlled reactions relates the recombination constant to the particle diffusivity (Smoluchowski) by
γTPA ) 4πDhT〈R〉
Figure 4. Current density vs voltage under pulsed conditions for PSFTAD devices (doped with 1% PtOEP) of varying thickness. The symbols correspond to thicknesses of (9) 717, (3) 590, (b) 525, (]) 398, and (1) 325 nm. The solid line is a fit to the data according to eq 6 for the 525-nm-thick device. The mobility is 6.5 × 10-5 cm2/(V s). The measurement was made in pulsed mode at 315 K. The inset shows the current density of the devices as a function of the film thickness on double logarithmic scales at ∼7.9 V. The slope of the fit is -2.9.
at 315 K. The room-temperature dc IV curves are very similar (not shown). From the pulsed IV curves (Figure 4), we extracted the hole mobility of PSF-TAD. Using eq 6, a fit of all IV curves yields an average hole mobility of (7 ( 2) × 10-5 cm2/(V s). This value compares favorably with the mobility obtained by timeof-flight (TOF) measurements in this material. At 318 K and a electric field of 2 × 105 V/cm, the TOF hole mobility is 2 × 10-4 cm2/(V s), only marginally depending on electric field.24 If one takes into account that the mobility in TOF is slightly overestimated because of the transit time determination,36 the agreement is very good. According to simulations, the average equilibrium mobility is roughly one-quarter of the TOF value.36 Now, we consider the dependency of the mobility on charge density37 to validate our approach for the application of eq 6. Assuming a dielectric constant, r, of 3, the charge density in the samples can be calculated. It ranges from 5 × 1014 cm-3 at 1 V for the 717-nm-thick device to 2 × 1016 cm-3 for the 325nm-thick device at 10 V. Taking into account the energetic disorder, σ ≈ 107 meV, in the material obtained from the temperature dependence of the hole mobility in TOF measurements,24 the material is characterized by an energetic disorder parameter of σ/kT ≈ 4. At such a disorder parameter and the low charge densities considered here, the mobility is independent of the charge density.37,38 Triplet-Polaron Annihilation (TPA). Having established that the hole current in our device is determined by space charges and that eq 6 describes the IV characteristics with sufficient accuracy enables us to obtain a rate constant for TPA according to eq 5. This yields γTPA ) (4 ( 1) × 10-13 cm3/s as an average value for all samples computed from the slope of the phosphorescence quenching vs voltage. This agrees favorably with TPA constants for small-molecule devices doped with phosphorescent Ir(III) complexes,13 but it is about 4 orders of magnitude lower than those in anthracene crystals.22,23 This is due to the reduced mobility of holes in PSF-TAD, which is about 4 orders of magnitude lower. The TPA constant of a conjugated polyphenylene-vinylene is 1 order of magnitude lower than our value.39 In that work, however, the value was obtained rather indirectly at very low temperature. It is questionable as to whether it is possible to obtain SCLC at 4 K and
(7)
where DhT is the sum of the diffusion constant of triplet excitons and polarons, DhT ) (Dh + DT), and R is the interaction radius. The diffusion constant of triplets is not known, but is assumed to be 2 orders of magnitude lower than the diffusion constant of holes, as has been observed in molecular crystals.32 In a ladder-type polyphenylene, triplet diffusion has been measured by Reufer et al.40 to be on the order of 10-5 cm2/s. However, in that work, triplet diffusion in a dense film was investigated. In our sample, the triplets were strongly confined to PtOEP because of the low concentration used. Therefore, we assumed the triplet diffusion to be negligible compared with the diffusion of holes. Using Einstein’s relation Dh ) µhkT/e enables us to calculate the diffusion constant of holes. This transforms eq 7 into
µhkT 〈R〉 e
γTPA ) 4π
(8)
With the hole mobility of (7 ( 2) × 10-5 cm2/(V s) at a temperature of 315 K and the value of γTPA ) (4 ( 1) × 10-13 cm3/s, the interaction radius is R ≈ 2 × 10-10 m. The radius is very small, which implies that the triplet is confined to the Pt center of PtOEP. The interaction radius can be interpreted as a probability of recombination of holes diffusing to a triplet exciton strongly localized on PtOEP. The majority of holes transported via PSF-TAD approaching a triplet on PtOEP will not annihilate. Thus, TPA is not “diffusion-controlled” in a traditional sense. Because the mechanism of the reaction is based on electron exchange, the probability of hole and triplet exciton annihilation is apparently related to the molecular structure of PtOEP and PSF-TAD. The high anisotropy of PtOEP, and hence the orientation relative to the polymer, plays a role as well. Quantum chemical calculations and investigations of different triplet emitters would be necessary to gain more insight into the mechanism. A cautionary note is appropriate regarding the application of eq 7. It implies isotropic diffusion of charges to the localized triplet exciton. This is justified because the polymer chain in a film is not extended and the charge transport is controlled by interchain hopping events41 and the short distance where intrachain transport occurs. This is the reason for the relatively low mobility in polymers.24,41 In the literature, there are reports of higher intrachain hole mobility probed by time-resolved microwave conductivity.42 Nevertheless, even in solution, the charge transport is fast only on a length scale of a few nanometers.43 After that distance, the carrier has to overcome chemical or topological defects on the polymer chain that limit the carrier mobility. Unfortunately, the low solubility of PtOEP prevents investigations of the concentration dependence of TPA. At higher dopant concentrations, an increase of the rate is expected
Triplet-Polaron Quenching in Conjugated Polymers because the diffusivity of the triplets should increase significantly. Investigations of lower concentrations are currently not feasible, because of the limited signal-to-noise ratio. Moreover, no effect concerning triplet diffusion is expected because 1% is already sufficient to ensure localization of the triplet exciton. However, the phosphorescence quenching will apparently be more pronounced in thinner samples where the charge density is higher assuming SCLC. In a device with a triplet emitter of similar lifetime and thickness relevant for OLEDs, the chargeinduced quenching will be about 50%. Because the mobility increases with charge density quite dramatically,37 it is expected that the TPA rate constant increases as well. 3. Conclusions We have shown that triplet-polaron interactions can be studied using the lifetime of the phosphorescent emitter PtOEP as a probe in space-charge-limited devices. The TPA constant is on the order of 4 × 10-13 cm3/s limited by the interchain transport of charges. The SCLC mobility of (7 ( 2) × 10-5 cm2/(V s) is in accord with independent transport studies based on the TOF technique. One can speculate that the quenching will be much more efficient in polymers in which the emitter is polymerized into the backbone, because the on-chain mobility is orders of magnitude higher in polymers than in the bulk43 as long as the chains are extended. An open question arises concerning the structure that an emitter should have to reduce TPA in doped conjugated systems. Studies are under way to address TPA and TTA in a description of electrically driven devices. Currently, we are extending our investigations to more soluble Ir(III) complexes having shorter phosphorescence lifetimes and exhibiting higher efficiencies in OLEDs. This will allow us to relate the interaction radius to structural properties. Recently, it has been proposed that TPA plays a major role in the magnetoresistance of organic materials.19 Further experiments will show if TPA is influenced by magnetic fields. 4. Experimental Section The chemical structure of the polyspirobifluorene copolymer PSF-TAD under investigation is shown in the inset of Figure 1. The polymer was synthesized from 50% of the comonomer spirobifluorene and 50% of the comonomer triphenylene amine dimer (TAD) as described previously.44 The phosphorescent emitter doped at 1% by weight into PSF-TAD was PtOEP. The polymer and the phosphorescent dye were obtained from Merck KGaA and Porphyrins Systems, respectively. Unipolar devices were fabricated on structured and cleaned glass substrates covered with a 150-nm layer of indium tin oxide (ITO, Merck). As the hole-injection layer, Pedot/PF26 (H.C. Starck GmbH) was used. The work function of PEDOT/PF27 is higher than that of the conventional PEDOT/PSS (Baytron P, 4.9 eV) but depends on preparation conditions. Films of about 30-nm thickness after being dried at 150 °C for 5 min were obtained by spin coating. The subsequent device fabrication steps were completed under inert atmosphere conditions. Before spin coating of the organic layers, the substrates were annealed at 150 °C for 10 min. To promote hole injection into the polymer, a cross-linkable hole-transport layer (HTL) of a triphenylenediamine derivative (termed AUPD) on top of the PEDOT/PF layer was used. The oxidation potential of the molecule AUPD (0.2V vs Fc/Fc+)33,34 is suitable to allow ohmic injection of holes into the HOMO of the polymer matrix PSFTAD. The HOMO of the TAD unit is at -5.39 eV, and that of PSF is at -5.63 eV.24 To yield insoluble films of AUPD, the
J. Phys. Chem. B, Vol. 111, No. 42, 2007 12079 solution in toluene was mixed with an adequate amount (4% by weight) of a solution of the photoinitiator {4-[(2hydroxytetradecyl)oxyl]phenyl}phenyliodonium hexafluorantimonate (OPPI, Sigma-Aldrich) in toluene. The concentration and spin speed were adjusted to give film thicknesses of 30200 nm. Cross-linking was achieved by 10 s of illumination of the films with 366-nm UV light, 10 s of heating at 100 °C, and a baking step at 150 °C. The cross-linking procedure was described in detail previously.45,46 PSF-TAD solutions in toluene containing 1% PtOEP were filtered and spin coated on top of the AUPD layer. The concentration and spin speed were adjusted to obtain films varying in thickness from 250 to 700 nm. To complete the structure, a 20-nm gold cathode covered with a 100-nm silver layer was evaporated at a base pressure of 10-6 mbar from resistively heated sources, using a shadow mask defining the active area of the diodes (3 mm in diameter). Film thicknesses were measured with a Dektak3 surface profilometer. Absorption and emission spectra were recorded with a Cary 50 Bio (Varian) UV/vis spectrometer and a Varian Eclipse fluorescence spectrometer, respectively. For time-resolved emission measurements, the samples were placed in a nitrogen flow cryostat (Cryovac). Photoluminescence (PL) was excited with a Nd:YAG (Surelite I, Continuum) pumped dye laser (NARROWscan, Radiant Dyes) operated at a repetition rate of 10 Hz. The pulse duration was 4 ns. The excitation intensity was varied by neutral density filters and measured simultaneously with a thermoelectric detector (Scientech). The laser light was focused onto the device contact (3 mm in diameter). The PL was collected and focused onto the entrance slit of the monochromator and detected by an intensified gateable CCD camera (Roper Scientific). A grating with 150 lines/mm was employed to achieve a spectral resolution of about 1 nm. Typical PL spectra were averaged over 100 laser pulses. For the timeresolved measurements, the ICCD camera was operated in a gated mode in which the width of the detection window can be varied and simultaneously delayed with respect to optical excitation. The minimum width of the detection window was 2 ns (gate). For measurements of PtOEP phosphorescence, a typical detection window of 1-µs width with a minimum delay of 10 µs was used. Time-resolved measurements were made at slightly elevated temperatures (315 K) under a continuous flow of nitrogen. For measurements of TPA, a unipolar hole current was driven through the device by applying a pulsed voltage of variable amplitude synchronously with the excitation laser pulse. The pulser (Agilent 8114) was triggered to apply the voltage to the device ca. 120 µs before laser excitation (RC time constant
