J. Phys. Chem. B 2005, 109, 17047-17054
17047
Spectroscopy and Photophysics of Self-Organized Zinc Porphyrin Nanolayers. 3. Fluorescence Detected Magnetic Resonance of Triplet States Tjeerd Schaafsma,*,† Inbar Dag,†,‡ Rolf Sitters,§ Max Glasbeek,§ and Efrat Lifshitz| Laboratory of Biophysics, Department of Agrotechnology and Food Sciences, Wageningen UniVersity, Dreijenlaan 3, 6703 HA Wageningen, The Netherlands, Institute for Molecular Chemistry, Department of Chemistry, Faculty of Mathematics, Computer Science, Physics and Astronomy, UniVersity of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands, and Chemistry Department and Solid State Institute, Israel Institute of Technology, Technion City, Haifa 32000, Israel ReceiVed: February 16, 2005; In Final Form: May 19, 2005
Fluorescence detected magnetic resonance (FDMR) has been applied to ∼25-nm-thick porphyrin films, containing ordered domains of zinc tetra-(p-octylphenyl)-porphyrin (ZnTOPP) spin-coated onto quartz slides. Illuminating the films at 1.4 K with 457.9-nm light from a continuous wave Ar+ laser produces at least two different, Jahn-Teller-distorted, ZnTOPP triplet species, labeled i and ii. Microwave-induced magnetic resonance of i and ii in the absence or presence of an externally applied magnetic field affects the fluorescence intensity of ZnTOPP, thus allowing FDMR. For triplet species i, formed in films spin-coated from toluene solution, the zero-field splitting (ZFS) parameters were determined as |D| ) (316.9 ( 0.1) × 10-4 cm-1 and |E| ) (32.0 ( 0.5) × 10-4 cm-1. By exposure of the spin-coated films to chloroform vapor at room temperature, triplet i is converted into species ii, with |D| ) (295 ( 3) × 10-4 cm-1 and |E| ) (121 ( 3) × 10-4 cm-1. For the excited triplet state of ZnTOPP in a toluene glass, ZFS parameters with values of |D| ) (295 ( 1) × 10-4 cm-1 and |E| ) (91 ( 1) × 10-4 cm-1 are found. From a combined study of the FDMR- and microwave-induced fluorescence spectra, i and ii are identified as unligated and ligated ZnTOPP triplet species, respectively. From the asymmetrically shaped zero-field FDMR signals of i, we conclude that the local crystal field perturbations of the stacked molecules are anisotropic. The FDMR results of the ZnTOPP films are compared with those for a film of zinc tetraphenylporphyrin (ZnTPP), which lacks the octyl substituents, and therefore is nonordered. Upon illumination, the ZnTPP films contain only a single, ligated, triplet species with ZFS parameters very similar to those of ligated ZnTOPP. At ∼5 K, the lifetime of triplet i is considerably shortened compared to that of ZnTOPP in a glass at the same temperature.
Introduction The two preceding papers report and discuss results of experiments applying optical spectroscopy to zinc porphyrin solid films with thicknesses in the nanometer range (“nanolayers”). This paper focuses on the nature and properties of the photoexcited porphyrin triplet states in these films as studied by means of optically detected magnetic resonance (ODMR). In particular, we have applied microwave-induced fluorescence intensity measurements, i.e., fluorescence detected magnetic resonance (FDMR),1 in the absence of a magnetic field, to these films at 1.4 K. The ODMR technique and its theoretical background have been treated in several excellent reviews.2-8 Photophysical processes involving triplet states in thin solid films including inorganic,9 organic, and polymer films10 have frequently been investigated by zero-field ODMR and its highfield equivalent electron paramagnetic resonance (EPR).11 In such films, excited triplet states have been found for a variety of different species such as traps, defects, interfaces, and aggregates. Furthermore, the short intermolecular distances in * Author to whom correspondence should be addressed. Fax: +31-317482725. E-mail:
[email protected]. † Wageningen University. ‡ On leave from the Chemistry Department and Solid State Institute, Israel Institute of Technology. § University of Amsterdam. | Israel Institute of Technology.
films favor delocalization by excitons or excitation and energy transfer. Finally, due to their high surface-to-volume ratio, thin films may easily adsorb trace amounts of atmospheric gases, e.g., moisture or oxygen, resulting in surface traps for excitation and/or charge carriers.12 The results of this work clearly illustrate that FDMR spectroscopy, in the absence or presence of an external magnetic field,4,13 offers several strong advantages in the investigation of molecular triplet species in ultrathin organic solid films. Not only is the detection limit for triplets in the investigated films below a nanomole, but in addition, the nature and properties of the detected triplet species can be determined using its zerofield splitting (ZFS) values in combination with the kinetics of the three spin levels. In the present work, the frequencies of the various FDMR transitions are assigned by combining the FDMR spectra with the corresponding microwaveinduced fluorescence (MIF) spectra1 and comparing the MIF spectra with the fluorescence spectra reported in the preceding paper. FDMR spectroscopy can be applied to localized triplets as well as to triplet radical ion pairs. Since photogenerated electron-hole pairs in solid films may interact with fluorescent species in their immediate environment, FDMR may also be utilized to report on the fates of these charge carriers when the latter are formed in the triplet state. Thus, dissociation, trapping, and/or recombination processes may be followed provided they
10.1021/jp0580569 CCC: $30.25 © 2005 American Chemical Society Published on Web 08/18/2005
17048 J. Phys. Chem. B, Vol. 109, No. 36, 2005
Schaafsma et al. contains three nondegenerate spin levels with spin functions |x〉, |y〉, and |z〉 and zero-field energies 2,3
1 1 2 X) D+E Y) D-E Z)- D 3 3 3
(1a)
In eq 1a
〈
〉
r2 - 3z2 3 D ) g2β2 4 r5 Figure 1. Molecular structure of ZnTOPP and ZnTPP, where R ) -C8H17 and -H, respectively.
affect the fluorescence intensity. High-field FDMR transitions, with g ≈ 2, in photoconductive polymers10 and layered semiconductors14 have been discussed to involve such electronhole pairs.15 In this work, films with thicknesses in the nanometer range of two different porphyrins on a quartz substrate have been investigated, one with self-organizing properties, i.e., zinc tetrakis-(octylphenyl)-porphyrin (ZnTOPP) and, for comparison, one without these properties, i.e., zinc tetraphenylporphyrin (ZnTPP) (Figure 1). The level scheme of Figure 2 describes the electronic and spin levels of a metallo-porphyrin with D4h symmetry, including the findings reported in the first preceding paper. As a result of Jahn-Teller distortion,16,17 the lowest excited triplet state T1
〈 〉
3 y2 - x2 and E ) g2β2 4 r5
(1b)
π-π* triplet states of oblate-shaped molecules have D > 0,18 so the energy of the zero-field |z〉 spin level will be lowest. Also, the labeling of the molecular x- and y-axes is interchangeable, and therefore E can always be chosen to have a positive value. In eq 1b, the triplet g-tensor has been assumed to be isotropic (g ≈ 2), β is the electronic Bohr magneton, and the broken brackets represent the average over the electronic wave function of the triplet state. D is a measure for the average distance r between the electron spins, with components x, y * 0 and z ) 0, and E reflects the in-plane asymmetry of the spin distribution over the porphyrin macrocycle. Microwave transitions between any two spin levels of T1 generally cause a change of their steady-state populations, resulting in a change of the net decay rate of T1 to the groundstate S0 and therefore of the S1 f S0 fluorescence intensity.1,4 Understanding the transport, trapping, and conversion processes of excited states in the investigated films and their structures (cf. previous papers) is important for various applica-
Figure 2. Energy level scheme for ZnTOPP for various situations A-D. Note the different energy scale used for B-D as compared to A. (A) Singlet electronic levels S0, S1, S2, and lowest triplet level T1, including S2 exciton components, labeled x and y. The S1 and T1 levels consist of four almost degenerate electronic levels. (B) Zero-field spin levels of the T1 level of ZnTOPP assuming D4h symmetry, a crystal field ∆, and B B0 ) 0. For simplicity, only spin levels of the lowest two electronic levels of the fourfold degenerate manifold are indicated for T1. Because of D4h symmetry, the two spin levels |x〉 and |y〉 are twofold degenerate. (C) T1 spin levels of ZnTOPP in the presence of a Jahn-Teller distortion resulting from ∆ * 0 and B B0 ) 0. (D) Zeeman effect on T1 spin levels for B B0 // z. FDMR transitions in the absence and presence of an external magnetic field B B0 are indicated by dashed arrows
FDMR in Zinc Porphyrin Nanolayers
J. Phys. Chem. B, Vol. 109, No. 36, 2005 17049
tions. Intersystem crossing to triplet exciton states or localized triplet traps may profoundly affect this excitation transport. Films of self-organizing (metallo-)porphyrins19 and phthalocyanines20 containing alkyl or alkoxy substituents have been utilized as energy collecting “antennas” on top of a semiconductor substrate in dye-sensitized solar cells.21-24 These ordered systems show enhanced excitation transport.25 Ordered ZnTOPP films that have been in contact with chloroform turn out to contain two different Jahn-Tellerdistorted triplet species,16,17 assigned to porphyrins with and without a ligand, based on their FDMR and MIF spectra. For nonordered ZnTPP films, these spectra reveal the presence of ligated triplet species only, irrespective whether the films have been in contact with chloroform or not. Experimental Section Tetraphenylporphyrin free base (H2TPP) and tetra-(p-octylphenyl)-porphyrin free base (H2TOPP) were synthesized by condensation of pyrrole with benzaldehyde and 4-(n-octyl)benzaldehyde, respectively, following standard procedures.26 Zinc was inserted into the porphyrins by refluxing a solution of the free base porphyrin in dimethylformamide (DMF) in the presence of excess zinc(II) chloride.27 The porphyrins were purified by chromatography over silica gel (Merck) with chloroform (Merck, p.a.) as the eluent and estimated to be >99% pure by thin-layer chromatography, absorption, and fluorescence spectroscopy. Solutions (5 × 10-3 M) of ZnTOPP or ZnTPP in chloroform (Merck, p.a.) or toluene (Merck, spectroscopy grade) were spincoated at 2500 rpm onto ∼0.5-mm-thick quartz slides (Suprasil I, P&H Glass Technology), thus resulting in smooth solid films. Through the use of the optical densities of the films at 550 nm, their estimated thicknesses were 25 ( 5 nm. The amount of deposited porphyrin was also determined by dissolving the film in a known volume of toluene. Through the use of the extinction coefficient28 550 ) 22 000 L M cm-1, the calculated thickness was found to be in good agreement with the above-mentioned value and corresponds to ∼100 porphyrin layers for all investigated films. After spin-coating, all films were heated to 150 °C during 5 min to remove adherent solvent from the film. For FDMR experiments, small pieces were cut from the same porphyrin-covered quartz slide to fit into the microwave helix. All samples were kept in the dark until used in the experiments. Three different films, in the following denoted as 1-3, were prepared by spin-coating onto the quartz slide: 1, ZnTOPP from toluene solution; 2, ZnTOPP from chloroform solution; 3, ZnTPP from toluene solution. In some experiments, 1 was exposed to saturated chloroform vapor at room temperature in a closed vessel during 30 min. For comparison of ZFS values, ZnTOPP was also studied in a toluene glass. Fluorescence and FDMR spectra were recorded with the films submerged in a bath of pumped liquid helium, at 1.4 K, inside a homemade cryostat. The films were excited by ∼70-mW 457.9- or 514-nm radiation from a continuous wave Ar+ ion laser (Coherent CR-5). Emission light was detected using an S20 (EMI) photomultiplier after dispersion through a colored glass cutoff filter (OG 540) and a monochromator (Jobin-Yvon, model HR1000). Microwaves were applied using the setup described previously.8,29 The sample was mounted inside a slowwave helix coupled to a semirigid cable in the cryostat. Microwave radiation was generated by a sweep oscillator (HP8609B) using appropriate plug-in units (HP-86999B for 1004000 MHz and 8691B for 1000-2000 MHz) and Hughes 1177H
Figure 3. FDMR spectrum of 20 µM ZnTOPP in a toluene glass: T ) 1.4 K; excitation, Ar+ 514-nm laser line; fluorescence detection, 590 ( 5 nm; microwave frequency sweep rate, ∼4 MHz s-1.
or Varian VZC-6961/VTC-6160/VA-615M traveling wave tube amplifiers; the amplified microwave output (∼20 W) was coupled by coaxial lines to the semirigid cable. The microwaves were square-wave amplitude-modulated using a Hewlett-Packard PIN diode switch with a modulation frequency of 280 Hz. The microwave-induced fluorescence changes were detected using a lock-in amplifier (Princeton Applied Research, model HR-8) with the microwave modulation frequency as a reference. The output of the lock-in amplifier was digitized and further processed by a PC. FDMR spectra were recorded by sweeping the microwave frequency over a 100-1500 MHz range and monitoring the change in fluorescence intensity, induced by connecting two spin levels (or three spin levels in a double resonance (FDELDOR) experiment, vide infra) of the lowest porphyrin triplet state at resonance conditions. To investigate the connection between the different resonances and their associated MIF spectra, we recorded the dependence of the amplitude of each resonance on the fluorescence wavelength. This was done by amplitude modulation of the microwaves at a fixed resonance frequency, scanning the wavelength of the emitted light using a monochromator and applying phase-sensitive detection (Princeton AppliedResearch, model HR-8) of the fluorescence intensity at the microwave modulation frequency. Fluorescence detected electron double resonance (FDELDOR) experiments, again at zero external magnetic field, were performed using two different microwave sources, coupled to two concentric slow-wave helixes, surrounding the sample. One of the oscillators was kept at a constant microwave frequency, pumping one of the FDMR transitions. The second oscillator was swept over the frequency range of the other resonances. High-field FDMR experiments were done at a magnetic field strength of 1.0 T using a superconducting solenoid coil, placed in a liquid helium cryostat and pumped at 1.2 K. The 457.9nm line of an Ar+ ion laser (Coherent, model Innova 70) was used for excitation. Results Figures 3-7 show the FDMR spectra of ZnTOPP in a toluene glass and in variously prepared films, respectively, at 1.4 K. The glass spectrum (Figure 3) contains two transitions corresponding to an increase in the fluorescence intensity, the major
17050 J. Phys. Chem. B, Vol. 109, No. 36, 2005
Schaafsma et al.
Figure 4. FDMR spectrum of a ZnTOPP film spin-coated from toluene solution (1): detection, total fluorescence emission; other conditions as in Figure 3; the dashed line indicates zero FDMR intensity. Figure 6. FDMR spectrum of a ZnTOPP film (1) spin-coated from toluene solution and subsequently exposed to chloroform vapor during 30 min; other conditions as in Figure 3; the dashed line indicates zero FDMR intensity.
Figure 5. FDELDOR spectrum, 1046-MHz pumped, of a ZnTOPP film spin-coated from toluene solution (1), other conditions as in Figure 4.
one at 613 ( 1 MHz and a weaker resonance at 1160 ( 6 MHz, whereas the spectrum of 1, using the entire fluorescence emission for detection, has two major, broad transitions (also reflecting fluorescence intensity increases) with maxima at 854 ( 5 and 1046 ( 5 MHz, denoted I and II (Figure 4). Note the different frequencies and intensity distributions of the glass and film resonances, both with an full width at half-maximum (fwhm) of ∼50-100 MHz. The average of each pair of resonances (887 and 950 MHz, respectively) are representative of the D-value of ZnTOPP in toluene and 1. These D-values are typical for a porphyrin triplet.1,4,16 The notably asymmetric line shape of the film spectrum was unaffected by varying the instrumental settings, in particular when the microwave sweep rate was lowered or the sweep direction was reversed. No resonance was observed for 1 at the difference frequency of ∼190 MHz. Since the samples may contain more than one triplet species (see below), it is not a priori obvious whether the resonances I and II (Figure 4) belong to one and the same triplet species. Pumping resonance II of 1 in a fluorescence detected electron double resonance (FDELDOR) experiment resulted in a relatively narrow (∼20 MHz fwhm) resonance (denoted as I-II), appearing at the difference frequency of 192 MHz of resonances I and II, showing that both resonances belong to one and the same triplet. Exposure of 1 to chloroform vapor initially results in considerable narrowing of resonances I and II (Figure 6) and the appearance of two additional resonances: III, corresponding to a fluorescence increase, at 1248 MHz, and IV at ∼948 MHz,
Figure 7. FDMR spectrum of the ZnTOPP film of Figure 6 after prolonged exposure to chloroform vapor; other conditions as in Figure 6.
corresponding to a fluorescence decrease. For a different part of the same film, only resonance III was reproducibly found (cf. Figure 7), indicating that the composition of various parts of the films is not uniform. Although the films obviously are not homogeneous, measurements were quite reproducible, provided that the sample in the microwave helix was not moved. Also, the same transitions, but not necessarily with the same relative intensities, were found for several films, when treated under identical conditions. Prolonged exposure of the film to chloroform vapor always produced spectra with a single, major resonance III, indistinguishable from spectra of 2. A general finding is that when the resonances I, II, and IV are absent only resonance III remains with a large amplitude, apart from some spurious weak satellites (Figure 7). The FDMR spectrum of 3 (not shown) was detected using the full fluorescence emission. Invariably a single, strong resonance with a positive sign at 1262 MHz is found with a minor satellite at 1300 MHz. Contrary to 1 and 2, the FDMR spectra of 3 did not show any dependence on the sample treatment, e.g., exposure to chloroform. The MIF spectrum associated with resonance I has main peaks at 584 and 642 nm (Figure 8A), whereas the MIF spectrum associated with resonance III has peaks at 594 and 652 nm (Figure 8B). Both spectra closely agree with the fluorescence spectra of Figures 10A and 10C, respectively, in the first paper.
FDMR in Zinc Porphyrin Nanolayers
J. Phys. Chem. B, Vol. 109, No. 36, 2005 17051
Figure 9. High-field FDMR spectrum at 1.0 T of a ZnTOPP film spin-coated from chloroform solution. The horizontal axis has been converted from a frequency scale to that of X-band electron spin resonance (ESR). The dashed line indicates the baseline.
Figure 8. Microwave-induced fluorescence spectra, (A) associated with resonance II and (B) associated with resonance III.
From the response of the major ∆mS ) (1 X-band EPR transition to a 14-ns, 532-nm laser flash at T ≈ 5 K, it can be concluded that the triplet lifetime of ZnTOPP in films of type 1 has an upper limit of 1 ms,30 shortened by at least a factor of 10 as compared to that of a zinc porphyrin in an organic glass in the same temperature range.31 Discussion Both the FDMR spectra and the MIF spectra of 1-3 present clear evidence for the presence of porphyrin triplet states1,4,16 The nonzero E-value for each of the observed porphyrin triplet species reflects a nonaxial symmetry of ZnTOPP and ZnTPP in their lowest triplet states; the scheme of Figure 2 includes this result that is thought to arise from a static Jahn-Teller distortion of the porphyrin molecule with D4h symmetry.16,17 The E-value strongly depends on the crystal field ∆, imposed
by the environment of the triplet species, giving rise to quite different splittings between the resonance pairs as shown, for example, by comparing Figures 3 and 4. The large differences between the glass and film FDMR spectra strikingly illustrate the quite different triplet-state properties in both media. ZnTOPP in toluene glass has an E/D ratio of 0.31, close to its maximum possible value of 1/3 for a molecule with an oblate shape. By contrast, the same molecule in the film is much less distorted, as shown by the E/D ratio of 0.10. The shortened lifetime of the triplet state in film 1 as compared to that in the glass indicates the presence of additional triplet relaxation and/or decay channels in the film. The results of the FDELDOR experiment of 1 (Figure 5), when pumping at 1046 or 854 MHz, show that the resonances I and II in Figure 4 belong to one and the same triplet and can be assigned as its zero-field D + E and D - E transitions. Without pumping, the 2E resonance is not observed, due to almost equal steady-state populations of the |x〉 and |y〉 spin levels (Figure 2), as commonly observed for porphyrins.1,4,16 Pumping at the maximum of one of the resonances I or II (Figure 4) causes a population redistribution over the triplet spin levels, thus allowing the 2E resonance to be observed in FDELDOR (Figure 5). The values of the zero-field splitting parameters as concluded from the FDMR and FDELDOR experiments for 1 are included in Table 1. The transitions I and II in the FDMR spectrum of 1 (Figure 4) exhibit some interesting features. (i) The line shapes are quite asymmetric and roughly each other’s mirror image around the average frequency of 950 ( 7 MHz; we note that a similar mirror asymmetry has been reported for the D + E and D - E
TABLE 1: ZFS Parameters D and E (units 10-4 cm-1) of Zinc Porphyrin Triplet States in Films and Solid Solutions compound
medium
ZnTOPP
film (1) film (2) toluene glass
ZnTPP
film (3) benzene/1% pyridine toluene glass
Db
Eb
methodc
reference
T (K)
316.9 ( 0.1 295 ( 3 295 ( 1
32.0 ( 0.5 121 ( 3d 91 ( 1
A, B A A
this work this work this work
1.4 1.4 1.4
315 ( 5 308 ( 5 315 ( 5
105 ( 5d 96 ( 5 e
A A C
this work ref 38 ref 35
1.4 4.2 5
a D and E are both assumed to be greater than 0. b Errors have been estimated considering the uncertainties in the position of the resonance peaks and the frequency markers. c A, FDMR; B, FDELDOR; C, ESR. d Calculated using the data from the high-field spectrum (Figure 9). e Not determined.
17052 J. Phys. Chem. B, Vol. 109, No. 36, 2005 resonances observed for a dilute solution of scandium(III) octaethylporphyrin monomers and dimers in a low-temperature 3-methyl pentane glass.32 (ii) Both resonances are considerably broadened (fwhm ≈ 50-100 MHz) as compared to those of the same films exposed to chloroform vapor (fwhm ≈ 30 MHz) (cf. Figures 4 and 6). (iii) Apart from the splittings, also the relative amplitudes of both transitions of ZnTOPP in a film are quite different from those in a glass (cf. Figures 3 and 4). Since the asymmetric line shape of both resonances does not change upon varying the microwave frequency sweep rate and the resonance line shapes were found to be independent of the sweep direction, slow spin level decay as the cause of the tails can be excluded. Thus, the line shape asymmetry must originate from the intrinsic properties of the porphyrin triplet state itself. When the asymmetry of the FDMR transitions of the porphyrin triplet state in films is discussed, several possibilities may be considered. A first possibility to explain the asymmetry would be to ascribe the resonance line shapes to triplet excitons. In ordered domains of ZnTOPP films, the intermolecular plane-to-plane distance is short enough to allow for some π-π orbital overlap and thus, in principle, allowing for triplet exciton motions to occur. Brenner33 has shown for (quasi-)one-dimensional molecular crystals that D + E and D - E ODMR transitions of triplet excitons may indeed exhibit asymmetrical line shapes due to k-dispersion. Simulated spectra using Brenner’s equations indeed yield asymmetric line shapes for both the D + E and the D - E transitions. However, the simulations also show that these line shapes are not related by mirror symmetry. Moreover, mobile triplet excitons are unlikely to occur in the films since the porphyrin molecules in the lowest excited triplet state are Jahn-Teller systems16,17 and local crystal field perturbations resulting from Jahn-Teller active vibrations may readily outweigh the weak excitonic interactions (a few cm-1), causing the excitation to remain largely localized. Another possible cause for the asymmetric broadening of the zero-field triplet-state transitions of ZnTOPP in films may be inhomogeneous broadening of the spin levels of the lowest triplet state. The smaller E-value for the lowest triplet state of ZnTOPP in the film and the larger width of the D + E and D - E transitions compared to that for the porphyrin in the glass implies that the crystal field perturbation exerted by the environment that leads to a nonaxially symmetric spin distribution is weaker and has a wider distribution for the porphyrin in the film as compared to that in the glass (cf. Figures 3 and 4). The above-mentioned crystal field perturbations are induced by imperfections in the environment of the stacked molecules and give rise to an inhomogeneous distribution of E-values. The influence of inhomogeneously distributed local fields is onesided or asymmetric in the sense that only larger E-splittings (with no smaller E-splitting values as counterparts) are obtained. Referring to Figure 2C, we thus propose that inhomogeneous broadening leads to a spreading of the spin level X to higher energies and of spin level Y to lower energies. The distribution function, characterizing this spreading, is assumed to be asymmetric, attaining its maximum for the smallest E-values. In this way, one anticipates asymmetrically shaped zero-field transitions such that the broadening for the D - E transition is toward its low-frequency side, whereas for the D + E transition the broadening is toward its high-frequency side. Thus, the model can qualitatively account for the experimental results. The narrowing effect on the spectra by exposure of the films to chloroform vapor (Figure 6) is attributed to a reduction of
Schaafsma et al. the spread of E-values albeit that the asymmetric line shape and their mirror symmetry are maintained. From the broadening of the zero-field transitions depicted in Figure 4, one might conclude that the ZnTOPP triplets in the stacks have a certain degree of in-plane disorder that appears to decrease upon exposure of the film to chloroform (Figure 6). Exposure of 1 to chloroform vapor also gives rise to a weak resonance III at 1248 MHz (Figure 6) of a second triplet species and a weak resonance IV of unknown origin (vide infra). It is noted that resonance III is the only one surviving a prolonged exposure to chloroform vapor (Figure 7). Also, its resonance frequency (1248 MHz) is not far from that of the D + E transition for the triplet state of ZnTOPP in a glass. In line with the discussion given above, we ascribe resonance III to strongly perturbed ZnTOPP molecules, probably ligated ZnTOPP, the ligand possibly being a small molecule such as water. The triplet species associated with resonance III is also found in film 2, which was spin-coated from chloroform solution. It can be clearly distinguished from the species with resonances I and II (cf. Figures 6 and 7). The MIF spectrum associated with resonance III (Figure 8B) identifies it as a ligated ZnTOPP monomer.34 Evidently, part of the nonligated traps become increasingly ligated by exposure to chloroform or its vapor; hence resonances I and II will disappear, and resonance III is the only resonance observed (Figure 7). Figures 6 and 7 demonstrate that the degree of ordering is not uniformly distributed over the film; i.e., some parts exhibit only the single resonance III, very similar to that in nonordered films of ZnTPP (see below). The MIF spectrum of signal III (Figure 8B) is red-shifted compared to that of signal II (Figure 8A). This agrees with the observed red shift of the fluorescence spectrum upon exposure to chloroform (cf. Figures 10A and 10C of the first paper), indicating that signal II (and as can be concluded from the FDELDOR experiment, also signal I) is due to the unligated monomer. Also, the Q(0,0) and Q(1,0) fluorescence wavelengths of 584 and 642 nm, found for the MIF spectrum of signal II in Figure 8A, indicate that ZnTOPP is nonligated, as reported values for the Q(0,0) wavelength change from 594 to 611 nm and for the Q(1,0) wavelength from 639 to 664 nm, respectively, upon ligation to pyridine.34 MIF spectra (not shown) obtained by modulated pumping of resonance IV are identical to those of resonances I and II. Therefore, this species is most likely also a nonligated monomer but with different ZFS parameters and an opposite sign of the fluorescence response to microwave resonance as compared to the first triplet species. The observation of only one zero-field resonance, III, for chloroform-treated ZnTOPP and ZnTPP films is remarkable. This situation is different from that for ZnTOPP in a toluene glass (Figure 3) and unligated ZnTOPP in a film (Figure 4) in which cases two resonances are observed. Generally, microwaveinduced transitions between the three spin levels as in Figure 2 are expected to result in 0, 2, or 3 FDMR resonances, unless E ≈ 0. We can exclude the latter possibility, considering the JahnTeller distortion of the porphyrin triplet state.16,17,32 Moreover, the calculated value for D of 416.3 × 10-4 cm-1 assuming E ≈ 0 is far outside the range typical for Zn porphyrin triplets. We may also exclude a low or vanishing microwave power in the frequency range around the |D - E| transition, in view of the results for ZnTOPP in a toluene glass (Figure 3). The high-field FDMR results (Figure 9) suggest the reason for the absence of the D - E and 2E resonances.. The outer Z+
FDMR in Zinc Porphyrin Nanolayers and Z- features in the high-field FDMR spectrum at 1.0 T are quite pronounced, but resonances in the XY regions are severely broadened and hardly detectable. The absence of the D - E and 2E resonances in Figure 7 and similarly for ZnTPP films can then be explained by assuming that their zero-field |y〉 spin level (Figure 2, column C) is severely broadened, so that only the D + E resonance is observed. The selective broadening of the |y〉 spin level is interesting, since it does not occur for ZnTOPP in toluene glass (Figure 3), suggesting that this selective broadening is related to the anisotropic environment of the porphyrin molecule in the film. The ZFS values D and E for 3ZnTOPP and 3ZnTPP from FDMR and high-field FDMR and EPR spectra, both from this and published work, are summarized in Table 1. As for chloroform-treated ZnTOPP films, a single resonance (not shown) at about the same frequency is also observed for nonordered ZnTPP films 3 (Table 1). The single resonance at 1254 MHz found for 3 is very close to that found for ZnTOPP after prolonged chloroform treatment (Figure 7), suggesting that the triplet species in both films are very similar. The presence of a different type of porphyrin, e.g., the free base, can be excluded as the cause of resonances III in 2 and 3 in view of their absorption, fluorescence, and MIF spectra. The 1254 MHz resonance of ZnTPP can be identified as the D + E transition of the ligated porphyrin monomer; this assignment is compatible with published ZFS parameters for the compound in a lowtemperature glass.35 Since the ZnTPP film was prepared using toluene as the solvent in the spin-coating procedure, the ligand could possibly involve a trace of water present in the toluene solvent. However, even when using sodium-dried toluene, resonance III survived for the ZnTPP film. Evidently, the ZnTPP molecules in the film appear as almost completely ligated, even when spin-coated from dry toluene. We tentatively ascribe the strikingly different ligation properties of ZnTPP and ZnTOPP films, both spin-coated from toluene, to the influence of the substitutional group -R, where -R is octyl for ZnTOPP and H for ZnTPP. In ZnTOPP films, the zinc center is protected from becoming ligated by small hydrophilic oxygen-containing ligands36 (e.g., water molecules37 from the atmosphere) by the octyl chains. In ZnTPP, the protecting influence of the hydrocarbon group is lacking, and the porphyrin may become ligated. However, by treatment of 1 with chloroform vapor or by spincoating ZnTOPP from a chloroform solution the hydrocarbon chains attached to the porphyrin are thought to be solubilized as a result of the high solubility of the hydrocarbon in chloroform. When the protective hydrocarbon barrier of ZnTOPP is eliminated by interaction with chloroform, the zinc center is now accessible to ligation, much in the same way as occurs for ZnTPP films. Conclusions From the results of this work, we conclude that: • At low temperatures, the triplet states of ZnTOPP and ZnTPP in films with thicknesses in the nanometer range have quite different photophysical properties as compared to those of the isolated compounds in solid glasses. • Conventional optical studies, reported in the preceding papers, and microwave-induced fluorescence studies (this paper) show that the ZnTOPP triplet state in films spin-coated from toluene has the spectral characteristics of a nonligated porphyrin; by prolonged exposure to chloroform vapor, the ZnTOPP molecules in these films become increasingly ligated, as shown by their triplet-state properties. • All observed triplet states exhibit a static Jahn-Teller distortion from fourfold symmetry, which is additionally
J. Phys. Chem. B, Vol. 109, No. 36, 2005 17053 perturbed by the crystal field exerted by the surrounding groundstate porphyrins. The asymmetric resonance line shapes of the |D + E| and |D - E| FDMR transitions can be explained by anisotropic, inhomogeneous broadening of the Jahn-Teller distortions. • The stacked ZnTOPP molecules in films exhibit localized triplet states, implying that the combined Jahn-Teller and crystal field interactions dominate the triplet exciton interaction. • In contrast to ZnTPP, the presence of hydrocarbon chains in ZnTOPP films prevents the access of small polar ligands, in particular, water molecules, to the zinc center of ZnTOPP. • Finally, the present work shows that FDMR and its complementary counterpart, microwave-induced fluorescence spectroscopy, may be successfully applied to sensitively and selectively study triplet states in nanolayers. Acknowledgment. This work has been supported as part of the research program “Organic Solar Cells; A Feasibility Study” by The Netherlands Organization for Energy and Environment Contract No. 146.100-024.4. T.J.S. gratefully acknowledges financial support from the Nata H. Sherman Lectureship Fund for research during a sabbatical stay at the Department of Chemistry, Technion, Haifa. The authors are indebted to Dr. T. J. Savenije for providing electron spin resonance data. Drs. H. Donker and W. van Schaik contributed to this work by useful comments. T.J.S. thanks Dr. J. Schmidt for illuminating discussions. References and Notes (1) Van Dorp, W. G.; Schaafsma, T. J.; Soma, M.; van der Waals, J. H. Chem. Phys. Lett. 1973, 21, 221. (2) van der Waals, J. H.; de Groot, M. S. In The Triplet State; Zahlan, A., Ed.; Cambridge University Press: New York, 1967; p 101. (3) Schmidt, J.; van der Waals, J. H. In Time Domain Electron Spin Resonance; Kevan, L., Schwarz, R. N., Eds.; John Wiley and Sons: New York, 1979; p 343. (4) Triplet State FDMR Spectroscopy; Techniques and Applications to Biophysical Systems; Clarke, R. H., Ed.; Wiley-Interscience: New York, 1982. (5) von Borczyskowski, C. Appl. Magn. Reson. 1991, 2, 349. (6) Kwiram, A. L. Annu. ReV. Biophys. Bioeng. 1982, 11, 223. (7) Biophysical Techniques in Photosynthesis; Amesz, J., Hoff, A. J., Eds.; Kluwer: Dordrecht, The Netherlands, 1996; Part 2, Chapter 17. (8) Glasbeek, M. Top. Curr. Chem. 2001, 213, 95-142. (9) (a) Svejda, P.; Maki, H. Chem. Phys. Lett. 1981, 83, 610. (b) Lifshitz, E.; Bykov, L.; Yassen, M. J. Phys. Chem. 1995, 99, 15262. (10) (a) Dyakonov, V.; Gauss, N.; Ro¨ssler, G.; Karg, S.; Riesz, W.; Schwoerer, M. Chem. Phys. 1994, 189, 687. (b) Lane, P. A.; Shinar, J. Phys. ReV. B 1995, 51, 10028. (c) Shinar, J. Synth. Met. 1996, 78, 277. (11) (a) Clarke, R. H.; Graham, D. J.; Hanlon, E. B. Photochem. Photobiol., Proc. Int. Conf. 1983, 2, 1011. (b) Lifshitz, E.; Kaplan, A.; Ehrenfreund, E.; Meissner, D. J. Phys. Chem. B 1998, 102, 967. (c) Kuroda, S.; Azumi, R.; Matsumoto, M.; King, L. G.; Crossley M. J. Thin Solid Films 1997, 295, 92 (12) (a) Kaufhold, J.; Hauffe, K. Ber. Bunsen-Ges. Phys. Chem. 1965, 69, 168. (b) Bo¨ttcher, H.; Fritz, T.; Wright, J. D. J. Mater. Chem. 1993, 3, 1187. (13) Cavenett, B. C. AdV. Phys. 1981, 30, 475. (14) Lifshitz, E.; Francis, A. H. Bull. Magn. Reson. 1989, 11, 403. (15) Sibley, S. P.; Francis, A. H.; Lifshitz, E. J. Phys. Chem. 1994, 98, 5089. (16) Chan, I. Y.; Van Dorp, W. G.; Schaafsma, T. J.; van der Waals, J. H. Mol. Phys. 1971, 22, 741. (17) Canters, G. W.; van Egmond, J.; Schaafsma, T. J.; van der Waals, J. H. Mol. Phys. 1972, 24, 1203. (18) Angiolillo, P. J.; Lin, V. S.-Y.; Vanderkooi, J. M.; Therien, M. J. J. Am. Chem. Soc. 1995, 117, 12514. (19) (a) Shimizu, Y.; Miya, M.; Nagata, A.; Ohta, K.; Matsumura, A.; Yamamoto, I.; Kusabayashi, S. Chem. Lett. 1991, 25. (b) Shimizu, Y.; Miya, M.; Nagata, A.; Ohta, K.; Yamamoto, I.; Kusabayashi, S. Liq. Cryst. 1993, 14, 795. (c) Kroon, J. M.; Koehorst, R. B. M.; Van Dijk, M.; Sanders, G. M.; Sudho¨lter, E. J. R. J. Mater. Chem. 1997, 7 (4), 615. (20) (a) Piechocki, C.; Simon, J. N. J. Chim. 1985, 9, 159. (b) Guillon, D.; Weber, P.; Skoulios, A.; Piechocki, C.; Simon, J. Mol. Cryst. Liq. Cryst.
17054 J. Phys. Chem. B, Vol. 109, No. 36, 2005 1985, 130, 223. (c) Ohta, K.; Jacquemin, L.; Sirlin, C.; Bosio, L.; Simon, J. New J. Chem. 1988, 12, 751. (21) Schaafsma, T. J. Sol. Energy Mater. Sol. Cells 1995, 38, 349. (22) Kroon, J. M.; Sudho¨lter, E. J. R.; Wienke, J.; Koehorst, R. B. M.; Savenije, T. J.; Schaafsma, T. J. Proceedings of the 13th European PhotoVoltaic Solar Energy Conference, Nice, France, Oct 1995; H. S. Stephens & Associates: Bedford, U.K., 1995; p 1295-1298. (23) Wienke, J.; Schaafsma; T. J.; Goossens, A. J. Phys. Chem. B 1999, 103, 2702. (24) Koehorst, R. B. M.; Boschloo, G. K.; Savenije, T. J.; Goossens, A.; Schaafsma, T. J. J. Phys. Chem. 2000, 104, 2371. (25) (a) Gregg, B. A.; Fox, M. A.; Bard, A. J. J. Phys. Chem. 1989, 93, 4227. (b) Markovitsi, D.; Le´cuyer, I.; Simon, J. J. Phys. Chem. 1991, 95, 3620. (c) Markovitsi, D.; Germain, A.; Millie´, P.; Le´cuyer, P. J. Phys. Chem. 1995, 99, 1005. (d) Lampre, I.; Markovitsi, D. J. Phys. Chem. 1996, 100, 10701. (e) Sigal, H.; Markovitsi, D.; Gallos, L. K.; Argyrakis, P. J. Phys. Chem. 1996, 100, 10999. (26) Adler, A. D.; Longo, F. R.; Kampas, F.; Kim, J. J. Inorg. Nucl. Chem. 1970, 22, 2443. (27) Adler, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.; Assour, J.; Korsakoff, L. J. Org. Chem. 1967, 32, 476.
Schaafsma et al. (28) Gouterman, M. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978; Vol. III, Chapter 1. (29) Glasbeek, M.; Hond, R. Phys. ReV. B 1981, 23, 4220. (30) Savenije, T. J. Personal communication. (31) (a) Miller, J. R.; Dorough, G. D. J. Am. Chem. Soc. 1952, 74, 2977. (b) Vogel, G. C.; Beckmann, B. A. Inorg. Chem. 1976, 15, 483. (32) Connors, R. E.; Leenstra, W. R. In Triplet State FDMR Spectroscopy; Techniques and Applications to Biophysical Systems; Clarke, R. H., Ed.; Wiley-Interscience: New York, 1982; Chapter 7, p 280. (33) Brenner, H. C. Annu. ReV. Biophys. Bioeng. 1982, 11, 241. (34) Humphrey-Baker, R.; Kalyanasundaram, K. J. Photochem. 1985, 31, 105. (35) van der Waals, J. H.; Van Dorp, W. G.; Schaafsma, T. J. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1979; Vol. IV, Chapter 5. (36) Nardo, J. V.; Dawson, J. H. Inorg. Chim. Acta 1986, 123, 9. (37) (a) Glick, M. D.; Cohen, G. H.; Hoard, J. L. J. Am. Chem. Soc. 1967, 89, 1966. (b) Scheidt, W. R.; Kastner, M. E.; Hatano, K. Inorg. Chem. 1978, 17, 706. (38) Benthem, L. Ph.D. Thesis, Wageningen Agricultural University, 1984.
