Excitation-Dependent Fluorescence to Phosphorescence Ratio of p -N

2500-W xenon arc lamp as the light source and a RCA 31034 photomultiplier in photon-counting mode as the detector.,O The resolution of both excitation...
0 downloads 0 Views 662KB Size
60

J. Phys. Chem. 1987, 91, 60-64

Excitation-Dependent Fluorescence to Phosphorescence Ratio of p -N ,N-Mmethylnitroanlline in Ethanol Alan J. Kallir,* Georg W. Suter, and Urs P. Wild Physical Chemistry Laboratory, Federal Institute of Technology, ETH- Zentrum, CH-8092 Zurich, Switzerland (Received: February 20, 1986; In Final Form: July 18. 1986)

The nature of the excitation-dependentluminescence of p-N,N-dimethylnitroaniline(DMNA) is examined. The results indicate that the anomalous luminescence behavior of DMNA is due to subtle differences in the molecular environment rather than a characteristic of the isolated molecule. The luminescent properties are ascribed to different H-bonded or rotameric forms, their energy levels being significantly determined by the local solvent polarity.

Introduction The spectroscopy of nitroanilines and their derivatives has gained attention for several reasons: These compounds belong to a class of highly polar aromatics D-Ar-A, D being an electron-donating and A an electron-accepting group, which show a very prominent and strongly solvent-dependent intramolecular charge-transfer band. The charge-transfer character of the different excited states has been discussed both in the singlet and the triplet manifold.1~2~'2 The strong solvent dependence of the absorption spectra results not only from the exceedingly large dipole moment difference of these molecules in the different electronic states but also from the formation of H bonds in protic solvents. Kamlet et aL3s4and Haberfield et aLs have shown that in fluid alcohols H-bond formation occurs to the nitro rather than to the amino group. Internal rotation of the substituents, in particular of the amino group, in the excited state leads to a complex luminescence behavior of these compounds. The relevance of this conformation for the fluorescence properties of aniline-like systems has amply been demonstrated in experimentalbL0and theoretical] studies. Dual fluorescence was found in many cases and the spectral components have been ascribed to emissions from species with different conformations of the amino group relative to the phenyl ring in the excited state, Le., planar and twisted (TICT) states6 Most of these studies, however, do not deal with phosphorescence. The most striking luminescence property of methylated nitro(DMNA), anilines, in particular of p-N,N-dimethylnitroaniline ratio on the excitation frequency. is the dependence of the aP/aF This effect has been observed also for a series of other aniline derivatives. The models that have been used to explain this unusual behavior may be divided into two groups, namely homogeneous models, which are based on the assumption that each individual chromophore in the sample shows an excitation-dependent emission spectrum, and inhomogeneous models, which ascribe the effect to the presence of different emitting species in the solution. In their pioneering papers on this subject, Khalil et al.L3r14 refer (1) Labhart, H.; Wagniere, G. Helu. Chim. Acta 1963, 46, 1314. (2) Wolleben, J.; Testa, A. C . J . Phys. Chem. 1977, 81, 429. (3) Kamlet, M. J.; Minesinger, R.; Kayser, E. G.; Aldridge, M. H.; Eastes, J. W. J . Org. Chem. 1971, 36, 3852. (4) Kamlet, M. J.; Kayser, E. G.; Eastes, J. W.; Gilligan, W. H. J . Am. Chem. SOC.1973, 95, 5210. (5) Haberfield, P.; Rosen, D.; Jasser, I . J . Am. Chem. SOC.1979, 101, 3196. (6) Grabowski, 2.R.; Rotkiewicz, K.; Siemiarczuk, A. J . Lumin. 1979, 18/19, 420. ( 7 ) Rettig, W.; Wermuth, G.; Lippert, E. Ber. Bunsen-Ges. Phys. Chem. 1979, 83, 692. (8) Wang, Y . ;Eisenthal, K. B. J . Chem. Phys. 1982, 77, 6076. (9) Rotkiewicz, K.; Rubaszewska, W. J . Lumin. 1982, 27, 221, (IO) Cazeau-Dubroca, C.; Peirigua, A,; Lyazidi, S. A,; Nouchi, G. Chem. Phys. Lett. 1983, 98, 511. (1 1) Lipinski, J.; Chojnacki, H.; Grabowski, Z. R.; Rotkiewicz, K. Chem. Phys. Lett. 1980, 70, 449. (12) Bigelow, R. W.; Freund, H.-J.; Dick, B. Theor. Chim. Acta 1983, 63, 177.

0022-3654/87/2091-0060$01.50/0

to the dual nature of the lowest absorption band of DMNA. These authors favor a homogeneous model in which the excitation deratio is ascribed to a very efficient inpendence of the aP/aF tersystem crossing (ISC) from S2 competing with internal conversion (IC) to SI. A theoretical study of this mechanism has been presented by Rashev.15 In connection with the solvent dependence of the fluorescence to phosphorescence ratio an inhomogeneous model has also been discussed,14 namely the formation of solute-solvent complexes which due to subtle changes in relative energies of the lowest excited states differ in their absorption spectra and show individual relaxation mechanisms, as discussed in a theoretical study by Plotnikov and Komarov.I6 A homogeneous mechanism, Le., efficient ISC from S, was also proposed by Dubroca et al.17518to explain the excitationratio of aniline and its N-methylated derivatives dependent aP/aF in non olar solvents. The effect was suggested to depend on the conformation of the molecule, Le., on the twisting angle between the amino group and the phenyl ring, which itself is influenced by solvent and temperature. On the other hand, Dobkowski et aLi9ascribed the excitation ratio in (dimethy1amino)benzaldehyde dependence of the aP/aF to the presence of different ground-state conformers (depending on the solvent and the temperature). They ruled out fast ISC processes from upper singlet states or the presence of H-bonded species in the solution. The aim of this work is to provide evidence for the heterogeneity of the DMNA luminescence in ethanol, which is responsible, at least partly, for the excitation-dependent ap/aFratio.

P

Experimental Section All luminescence experiments, including the acquisition of the two-dimensional spectra (TDS) were performed on a fully computer-controlled high-resolution spectrometer, consisting of two Spex 1402 monochromators equipped with an Osram XBO 2500-W xenon arc lamp as the light source and a RCA 31034 photomultiplier in photon-counting mode as the detector.,O The resolution of both excitation and emission was better than 1 nm. A front-face geometry was applied and a chopper, placed in the excitation beam, combined with electronic gates in the detection circuitry served as a phosphoroscope. It was operated at 100 Hz, Le., much faster than the phosphorescence decay, and thus both the pure phosphorescence and fluorescence spectra could be obtained concurrently. The intensities were not corrected for the (13) Khalil, 0. S.; Seliskar, C. J.; McGlynn, S. P. J . Chem. Phys. 1973, 58, 1607. (14) Khalil, 0. S.; McGlynn, S. P. J . Lumin. 1975/76, I I , 185. (15) Rashev, S. J . Lumin. 1981, 23, 305. (16) Plotnikov, V. G.; Komarov, V. M. Spectrosc. Lett. 1976, 9, 265. (17) Dubroca, C.; Lozano, P. Chem. Phys. Lett. 1974, 24, 49. (18) Cazeau-Dubroca, C. J . Lumin. 1984, 29, 349. (19) Dobkowski, J.; Kirkor-Kaminska, E.; Koput, J.; Siemiarczuk, A . J . Lumin. 1982, 27, 339. (20) Suter, G. W.; Kallir, A . J.; Wild, U. P. Chimia 1983, 37, 413.

0 1987 American Chemical Society

Fluorescence to Phosphorescence Ratio of D M N A spectral response of the detection or excitation subsystems. Therefore, the @'p/@F ratios discussed in the following sections are given in arbitrary units. They may, however, be compared to each other, since they were evaluated from spectra recorded with the same apparatus under identical conditions. For the acquisition of the phosphorescence decay curves, a mechanical shutter was used to cut off the excitation light. The decay curve was recorded thereafter by counting the photons within a sequence of 1024 intervals using a multichannel scaler (MCS).*I The time-resolved emission spectra (TRES) were obtained by measuring the phosphorescence decay at different emission wavelengths. The data obtained from such an experiment were fitted to

t "

24000

i

where Ai(Iem)is the emission spectrum and T~ the lifetime of the ith component. These data are conveniently analyzed in the following manner. First, the data are summed along the emission axis, significantly improving the signal-to-noise ratio of the temporal information compared to an individual analysis a t each emission point, and subsequently fitted with I(?) =

The Journal of Physical Chemistry, Vol. 91, hro. 1, 1987 61

EEi exp(-t/Ti) i

to provide good estimates of the T~ via a nonlinear least-squares procedure. These T~may then be inserted into eq 1 and the Ai&,) may be easily obtained, since eq 1 is now linear, via a linearleast-squares fit at each I , point. All the phosphorescence TRES data reported in this paper were analyzed with this method with matrices with 128 points in I,,, by 128 points in time. The fluorescence decays were recorded on a single-photontiming apparatus with a mode-locked argon ion laser in combination with a synchronously pumped cavity-dumped rhodamine 6G dye laser as the excitation source. The dye laser output was frequency doubled with an angle-tuned crystal, giving picosecond pulses at 301 nm. This single-photon-timing fluorimeter has been described in detail elsewhere.22 Fluorescence time-resolved emission spectra were measured by observing the fluorescence decay at different emission wavelengths. The problem of deconvoluting the data with the excitation pulse was avoided by discarding the first 1.35 ns of each decay. The remaining data was assumed to be free of contamination from the excitation pulse and could be analyzed with the method described above. The and 44 in time. matrices here were smaller, 40 in wavelength (km) Ground-state recovery experiments were conducted with the 365-nm line of an argon ion laser (Coherent Radiation Model CR52). The laser beam was split into two, a "populate", I,, and an "inquiry", ling, beam, with a 998:2 intensity ratio. The populate beam could be chopped with a mechanical shutter. In these experiments the full intensity was used to populate the triplet state via the singlet state. When the shutter for the populate beam was closed, the time-dependent ground-state repopulation could be observed with the MCS described above by using the fluorescence at the frequency, I O M , provided from excitation with the inquiry beam. Since the inquiry beam was very weak compared to the populate beam there was only an insignificant and time-independent ground-state depopulation due to the inquiry beam. Great care was taken to ensure that both beams were collinear. Alternatively, the excitation part of the spectrometer was used as the inquiry beam, enabling the ground-state recovery to be observed with a frequency-dependent inquiry beam. This proved to be most useful (see Results), although a good overlap was more difficult to achieve. An Osram XBO 150-W xenon lamp in combination with an optical band-pass filter transmitting light between 350 and 450 nm was used for most irradiation experiments. Other irradiation (21) Tschagaelar, __ R.; Kallir, A. J.; Forrer, J.; Wild, U. P., manuscript in preparation. (22) Canonica, S.; Wild, U. P. Anal. Instrum. (N.Y.) 1985, 14, 331. (23) Borkowski, W. L.; Wagner, E. C. J . Org. Chem. 1952, 17, 1128.

22000

20000

20000

18000

EMISSION I C M - 1 I Figure 1. Two-dimensional spectrum (TDS) of DMNA in EtOH (99.6%) at 77 K.

experiments were carried out with the UV lines of an argon ion laser. Again, special care was taken to ensure a good spatial overlap between the illuminated sample volumes in the irradiation and the subsequent luminescence experiments. All absorption experiments were performed on a Beckman ACTA V spectrometer equipped with a modified sample compartment, which allowed a cryostat to be inserted and irradiation of the samples without removing them from the cooling chamber. For the 77 K experiments the samples were placed in an Oxford DN704 cryostat, whereas below this temperature an Oxford CF204 cryostat was used. n-Hexane (UV-grade), n-octane (puriss), and n-nonane (purum), all from Fluka, were dried over metallic sodium and filtered over aluminum oxide before use. Ethanol of optical quality was carefully distilled and fractions with transmission greater than 90% (220 nm, 1 cm) were taken. Absolute ethanol was used as obtained from Fluka. DMNA was synthesized by the method of Wallach described in ref 23 and purified to 99.9% (GC) via sublimation and recrystallization. Subsequently it was recrystallized another 3 times from dioxane. Diethyl ether, dioxane, acetone, tetrahydrofuran, (all UV grade) and dibutyl ether were used as supplied by Fluka.

Results General Luminescence Properties. In aliphatic hydrocarbons, Le., nonpolar solvents, DMNA is essentially nonfluorescent but it shows a rather strong and well-structured phosphorescence. In Shpol'skii matrices of n-hexane, n-octane, and n-nonane at 4.2 K quasi-linear phosphorescence spectra were obtained with a very complicated site structure. The lack of fluorescence obviously prevented the observation of the excitation-dependent @.P/@'F ratio in nonpolar environments. With increasing polarity of the solvent fluorescence appears and the phosphorescence is red-shifted and becomes broad and stru~tureless.~~ Unfortunately, none of the polar Shpol'skii solvents tried (such as diethyl ether, dioxane, acetone, and tetrahydrofuran) gave quasi-linear spectra at 4.2 K. These solvents show both unstructured fluorescence and phosphorescence. Attempts to make a polar but ordered solvent, with a similar structure to the Shpol'skii matricies, were unsuccessful. A 99:l mixture (v/v) of n-nonane and dibutyl ether a t 4.2 K gave an unstructured fluorescence, whereas the phosphorescence consisted of a quasilinear component on top of a broad and unstructured spectrum. This result could be well explained with the notion that the luminescence came from DMNA occupying two different environments in the mixture solvent, a polar region, leading to the unstructured fluorescence and phosphorescence, and a nonpolar region, leading to the structured phosphorescence. The lack of structured fluorescence and phosphorescence prevented a highratio. resolution study of the aP/aF In ethanol (EtOH) the fluorescence and phosphorescence emissions are of comparable intensity. Excitation at 27 770 cm-' gives fluorescence of two bands of nearly equal intensity, at 22 620 and 21 670 cm-', whereas the phosphorescence maximum is at

62 The Journal of Physical Chemistry, Vol. 91, No. I, 1987

Kallir et al.

18 900 cm-’ with a distinct shoulder at 18 025 cm-I. A The excitation dependence of the luminescence of DMNA in EtOH at 77 K is visualized in the TDS shown in Figure 1. The f = 165215 MS excitation maxima of the fluorescence and the phosphorescence are at 23 260 and 26 210 cm-’, respectively. In addition to the edge excitation red shift reported for the phosphore~cence,’~ a z 7 = 42Qt L MS similar shift is observed also in fluorescence. Such shifts are indicative of strong interactions of the chromophores with the solvent and of solvent reorientation times that are of the order 20000 18000 16000 of the excited-state lifetime.24 WAVENUMBERS [CM? In agreement with ref 13 no concentration dependence of the L20 LM LBO 510 aP/aF ratio on the concentration in the range between lo-’ and lo-) M was found for a given excitation frequency. In addition this ratio did not depend on the cooling rate or on the oxygen content of the sample. > Changing the water concentration in the EtOH affects not only 1 6 t 0 2 NS’ ratio but also the spectral shape of the fluorescence the aP/aF bands. The aP/aF ratio in EtOH containing 0.2%H 2 0 is a factor INTENSITY 2 1 (SHORT LONG1 1.5 larger than that observed for EtOH containing more than 5% W z H20. The maximum intensity of the fluorescence peak at 22 620 cm-’ decreases by about 20% relative to the maximum intensity of the peak at 21760 cm-’, whereas the shape of the phosphorescence remains essentially unchanged upon the addition of 10% H2O to 99.8% EtOH. The aP/aFratio depends on temperature. Below 26 K no changes occur, whereas above this temperature a distinct increase of aP/aF ratio is observed. This increase is clearly excitation 420 60 L80 510 dependent: Excitation a t 27 780 cm-l increases the @ p / a F value WAVELENGTH [NMI by a factor of 1.75 between 26 and 77 K, whereas upon excitation Figure 2. Time-resolved emission spectra (TRES) of DMNA in EtOH at 24 600 cm-’ the factor is only 1.4. (a) in phosphorescence at 77 K and in fluorescence at (b) 10 K and (c) Phosphorescence Lifetimes and Time-Resolved Spectra. The 77 K. phosphorescence decay of DMNA in EtOH at 77 K excited a t 27 770 and 23 800 cm-’ may be well fitted with a double-expowhere I p is the intensity of the populate beam, I,, is the frenential function. The lifetimes of the two components are 420 quency ofthe populate beam, Iinqis the frequency of the inquiry f 4 and 165 f 15 ms. These lifetimes may be considered to be is the frequency at which the fluorescence is beam, and 3oobsd in agreement with the lifetime obtained by KhalilI3 of 350 f 90 observed. ms since the steady-state intensities are of the order of 1O:l. The The characteristic time, T , obtained from these fits was 425 spectral plot of the amplitudes obtained from the lifetime analysis f 50 ms; i.e., it is in good agreement with the longer lifetime of a TRES is shown in Figure 2a. The spectrum of the short-lived belonging to the dominant phosphorescence decay. component is somewhat less structured and red-shifted with respect In a uniform sample LY should not depend on the frequency of to the spectrum of the longer lived component. the inquiry beam or on the observation frequency, provided that Upon excitation into the red edge of the phosphorescence exthis is not in the phosphorescence band. The spectral dependence citation spectrum, at 23 000 cm-’, the phosphorescence intensity of a therefore provides a sensitive test on the “uniformity” of the is too small for a proper double-exponential analysis. A singlefluorescence emission. From the plot (Figure 3a) it is apparent exponential analysis yields a lifetime of 366 ms, which may be that LY depends on inquiry frequency in a way similar to the aP/aF understood as in ref 13 as an average of the two lifetimes reported ratio.13 above. However, an analysis using the lifetimes above, as fixed One should note that a(tinq)is also sensitive to photochemical parameters, gave a significantly larger contribution of the short changes in the sample (vide supra), which might well occur due component: Le., the relative steady-state intensity of the short-lived to the relatively large intensity of the pumping light in the recomponent is larger upon red-edge excitation. In fact, the population experiments. Since repeated scans with the same phosphorescence spectrum obtained when exciting at 23 000 cm-’ sample yielded the same results, the effect of photochemistry may is similar to the subspectrum of the short component: namely, be ruled out. red-shifted and weakly structured. Upon cooling the short-lived Effect of Irradiation. Changes in the ground-state populations component disappears and the lifetime of the dominant component of composite samples have been induced photo~hemically,~~ and significantly increases to about 480 f 4 ms. ratio of DMNA changes upon intense radiation. For the aP/aF Fluorescence Lifetimes. The time-resolved fluorescence spectra example, if a sample is irradiated with the UV lines of a argon are shown in Figure 2b. At both 10 and 77 K the fluorescence ion laser there is a decrease in the emission intensity. The emission may be described as a double exponential. The longer lived and spectra observed in a laser-burning experiment are shown in Figure less intense spectral component, with a lifetime of about 4 ns, is 4. The spectra renormalized to equal intensity at 22 550 cm-’ red-shifted and somewhat less structured than the shorter lived clearly show that the phosphorescence intensity decreases relative spectral component, with a lifetime of about 1 ns. to the fluorescence. Ground-State Repopulation Experiments. In order to deterThe effect of irradiation with intense light from a xenon lamp mine if the fluorescence and the phosphorescence of DMNA in on the spectrum of DMNA in EtOH at 77 K is similar. A broad EtOH originate from identical molecules, repopulation experiments band range from 350 to 450 nm was selected. Both the were carried out. fluorescence and the phosphorescence intensities are lowered by The experimental curves, obtained with laser “populate” and but the decrease in the phosphorescenceis definitely “inquiry” beams, observed at a wide range of frequencies, E ~ ~ the ~ irradiation, ~ , larger than that in the fluorescence. Note that the spectral shape could be well fitted to a function of the type of the phosphorescence and fluorescence remain essentially unI(t,~poQtnq,~od = I ( m ) [ l - a(Zpop,Ipop,tinq) exp(-t/r)l (3) I

(24) Itoh, K.; Azumi, T. J . Phys. Chem. 1975, 62, 3431.

(25) Suter, G. W.; Wild, U. P.; Schaffner, K. J . Phys. Chem. 1986, 90, 2358.

The Journal ofphysical Chemistry, Vol. 91, No. I , 1987 63

Fluorescence to Phosphorescence Ratio of DMNA A

L

Z

2 c

a U

0

J

I 27000

25000

m

m

23000

a

WAVENUMBERS [CM”] 1

I

I

0

0.35

0,1301-

350

0.25

LOO

L50

WAVELENGTH [NMl

Figure 5. Changing absorption spectra upon strong irradiation (the irradiation time is in minutes; see text).

28000

26000

2LOOO

WAV EN UMBE R S [ CM-’I Figure 3. (a) Ground-state recovery factor, a,as a function of inquiry ratio after irrafrequency (see text), tow= 21 670 c d . (b) Ip/IF(Bex) diation divided by the Ip/IF(B,,)ratio before strong irradiation with a 150-W Xe lamp for 20 min. Curve 1: fluorescence B,m = 21 740 cm-l and phosphorescence Bow = 19050 cm-I. Curve 2: fluorescence ?ow= 22 730 cm-I and phosphorescence Bow = 19 050 cm-’.

T = O MIN

BURN AT 2 7 6 0 0 CM-’ I365 N M ) 77 K . 9 6 % ETOH

23000

20000

17000

WAVENUMBERS [CM-’]

F g u e 4. Effect of laser irradiation on the emission spectrum of DMNA in EtOH (96%) at 77 K. The absolute intensities are in the ratio of

2:lS:l. The aP/OF ratio decreases upon irradiation,which is inconsistent with a homogeneous model yet may be explained by the presence of different DMNA species with different and individual ap/aFratios. Those with relatively high aP/aFratios degrade preferentially. changed. Figure 3b shows the ratio of the ZP/Z&J after and before irradiation. Note that ZP/ZF(S~~)does not decrease uniformly; a larger decrease may be observed in the red edge, at about

24500 cm-I. Also noteworthy is the fact that the ZP/ZF(bex) function does not show the “step” that is reported by Khalil.I3 The photochemical changes may also be observed in the absorption spectrum. Besides a distinct decrease in the entire absorption band, a clear change in its spectral shape may be observed: The decrease is smaller at the absorption shoulder at 24 460 cm-’ (408.8 nm) than at the absorption maximum at 25 250 cm-’ (396.1 nm). This together with a new absorption in the red edge of the absorption band results in a shift of the absorption maximum from 25 250 to 24 560 cm-’ (415.7 nm). Thus, at least some of the photoproducts have an absorption spectrum that strongly overlaps with the absorption spectrum of a fresh solution. The relevant spectra are shown in Figure 5.

Conclusion In summary the results for ethanol show the following: (1) a concentration, cooling rate, and oxygen content independent aP/aF ratio, in agreement with Khalil;I3 (2) a solvent and H,O concentration-dependent aP/aF ratio, indicating that the environment plays a role, and a temperature-dependent aP/aFratio, in agreement with the expectations of Plotnikov;16 (3) an excitaratio, implying either a multition-frequency-dependent aP/@F component sample or a wavelength-dependent branching ratio between the I C and ISC rates; (4) a double-exponential phosphorescence decay, with distinct spectra associated with each lifetime, Le., a nonuniform phosphorescence; (5) a ground-state repopulation observed in fluorescence with essentially the same time constant as the principle phosphorescence decay (thus at least part of the fluorescence and phosphorescence are kinetically related); (6) a time-dependent fluorescence spectra and a inquiryfrequency-dependent a factor in the repopulation experiment, indicating a nonuniform fluorescence; (7) a change in the aP/aF ratio upon irradiation of the sample both with a narrow and a broad band light source; (8) a change in the shape of ZP/ZF(t,,) upon irradiation, indicating that either the fluorescence or the phosphorescence or both are nonuniform. In particular, this rules out the possibility of two species, one of which is only fluorescent and the other of which is only phosphorescent. When taken individually none of the results is conclusive with respect to the origin of the excitation-dependent emission of D M N A in EtOH. However, as an ensemble considered simultaneously, the only explanation that fits our experimental results is that the excitation-dependent emission of DMNA in EtOH is due to nonuniformities in the sample. Aggregates must be ruled out, in agreement with ref 13, because the effect is independent of concentration. Thus, it is beyond all resonable doubt that there is more than one emitting species that contributes to the observed

J. Phys. Chem. 1987, 91, 64-68

64

effects. In principle, the nonuniformity may be due to impurities in the sample (either impurities in the DMNA itself or impurities due to chemical or photochemical reactions in the sample solutions). We strongly discount this explanation, since excitationratios have been reported not only for DMNA dependent aP/aF but also for a series of similar systems by many other aut h o r ~ . ~ - ~ ~Moreover, , ~ ~ , ~we ~ ,have ~ ~ thoroughly - ~ ~ checked the purity of our sample, and it is extremely unlikely that all the authors (vide supra) used impure samples. This indicates that the excitation-dependent @p/@F ratio is an inherent property of anilines rather than an artifact. We construct the following hypotheses in an attempt to rationalize the observed phenomena. Hypothesis 1: H-Bond Formation. In principle, the DMNA molecule may form two different types of H bonds to the protic solvent, i.e., to either the amino or the nitro group. The results obtained for fluid alcohol^^-^ would suggest H-bond formation to the nitro group. However, H-bond formation to the amino group has also been proposed10 for rigid matrices, resulting in a twisted ground state. Note that dynamic studies would suggest that at room temperature a partially formed ground-state H bond may be strengthened in the excited state, resulting in a geometry change and giving rise to a dual fluorescence.8 Thus, a number of different H-bonded conformers of the DMNA molecule may be present in our samples. The fact that the @pp/@Fratio changes with H 2 0 concentration speaks for a H-bonding effect. Further, ratio and the phosthe temperature dependence of the aP/QF phorescence lifetimes may be explained by considering that interethanol H bonding (chain building) is temperature dependent. With this hypothesis, the TRES result from different H-bonded species. The photochemistry could be explained by small dif-

ferences in at least one of the photophysical parameters, absorption spectra, quantum yields, ISC rate, and IC rate of the H-bonded and free forms. Hypothesis 2: Rotamers. The DMNA molecule has two possible rotation coordinates: the amino-phenyl and the nitrophenyl axes. Fundamentally, a distribution of rotation angles is always present, and freezing the solvent may fix this distribution. Naturally, the different rotamers possess different energetics, and the excitation-dependent @p/@'F ratio may be explained by different excitation spectra, lifetimes, ISC rates, IC rates, and quantum yields for the different r o t a m e r ~ . 'In ~ this regime, the TRES results represent "averages" of different rotameric classes, and the temperature dependence of the @p/@F ratio and the phosphorescence lifetimes could be explained by changing the populations of the rotamers. Indeed, the H 2 0 dependence may be considered to alter the polarity of the solvent and thereby influence the energetics of the rotamers and their distribution. Once again, the photochemistry could be explained by small differences in the molecular parameters (vide supra). Moreover, we would like to point out that a combination effect of different ground-state rotamers due to H bonding and H-bond formation in the excited state must also be considered. Notwithstanding these hypotheses our results indicate that the @p/@F ratio is not a characteristic of the isolated molecule but must be considered to be a characteristic of the different molecule-solvent environments.

Acknowledgment. We thank the Swiss National Science Foundation for financial support. Registry No. DMNA, 100-23-2.

The Influence of Magnetic Fields on the Photochemical Reactivity of Coordination Complexes: Photophysics and Photochemistry of Chromium( I I I ) Polypyridine and Hexacyanocobaltate(I I I ) Complexes G. Ferraudi,* G. A. Argiiello, and M. E. Frink Radiation Laboratory, University of Notre Dame, Notre Dame, Indiana 46556 (Received: July 1 1 , 1986)

Intense magnetic fields, H > 10 kG, induce changes in the absorption and emission spectra of the long-lived doublet excited states in Cr(II1) polypyridine complexes and increase the rate of excited-state relaxation by ca. 10%. Similar effects were detected with polycrystalline samples of K,[CO(CN)~]between -35 and -130 O C , while a 100% increase of the quantum yield for the cyano aquation was measured in room temperature photolyses under a 24-kG field. The effect of the field on the photophysics and photochemistry of the complexes is discussed in terms of a mechanism that involves the coupling of the field to the excited-state electronic angular momentum and the effect of such a coupling on the transition probabilities.

Introduction A number of studies about magnetic field effects on photochemistry have been reported in recent years.'-* To a large proportion, these reports dealt with the perturbation of the radical pairs reactivity by weak magnetic fields, e.g. H I 1 kG.1-3 That intense magnetic fields, e.g., H > 1 kG, affect the electronic levels of polyatomic molecules is known from the observation of the Zeeman effect with organic and inorganic compounds, the

magnetic-field-induced fluorescence quenching in gas phase, and the influence of magnetic field on the rate of triplet-triplet annihilation reaction^.^-^ Moreover we have recently shown that such intense fields are able to affect the photosubstitution reactions of coordination complexes, Le., the ammonia and acido photoaquation reactions, eq 1 and 2, of Rh(NH3)5X2+,X = C1 or Br.'

(1) Tanimoto, Y.; Udagawa, H.; Katsuda, Y.; Itoh, M. J. Phys. Chem.

1983. 87. 3916. - ,.. .

Turro, J. N.; Chow, M.-F.; Chung, C.-J.; Tanimoto, Y.; Weed, G. C.

. Chem. SOC.1981, 103, 4574.

Scaiano, J. C.; Abuin, E. B. Chem. Phys. Lett. 1981, 81, 209. Atkins, P. W.; Stannard, P. R. Chem. Phys. Lett. 1977, 47, 113. Matsuzaki, A.; Nagakura, S. J. Lumin. 1979, 18, 115. Douglas, E. A. Can. J . Phys. 1958, 36, 147. Ferraudi, G.; Pacheco, M. Chem. Phys. Lett. 1984, 112, 187. Frink, M. E.; Geiger, D. K.; Ferraudi, G. J. J. Phys. Chem. 1986, 90,

0022-3654/87/2091-0064$01.50/0

We have investigated in this work the magnetic perturbation of the photophysical and photochemical transformations of ligand (9) Chen,W.-H.; Rieckhoff, K. E.; Voigt, E. M. Chem. Phys. 1985, 95, 123.

0 1987 American Chemical Society