Photoisomerization

The construction of dynamic molecuar machines is a major challenge for chemists, physicists, and materials scientists. Our group, in collaboration with synthetic chemists, is characterizing the photoisomerization of a molecule that is the precursor for our 'light-driven molecular motor'. Light-mediated interconversion of Z and E stereoisomers of 1,2-diaryletheanes (e.g., trans- and cis-stilbene) has been studied extensively, but the photochemistry of 1,1-diarylethenes has been less well investigated. We are interested in the dibenzofulvene (9-methylenefluorene) chromophore as a photon-activated drive unit for molecular devices, including molecular motors. The thermal barriers for rotation about the exocyclic double bond of fulvene has been theoretically modeled, and experimentally determined in the case of substituted fulvenes. Excited state calculations suggest that fulvenes should photoisomerize, but this has not yet been shown experimentally. Thermal isomerization of 2-nitro- and 2-bromo-9-(4-nitro-benzylidene)fluorene was reported more than five decades ago. The molecule of interest for our group is (2-t-butyl-9-(2,2,2-tri-phenylethylidene)-fluorene). This molecule has two isomeric configurations, Z (cis isomer) and E (trans isomer) that undergo isomerization upon the absorption of a photon. This isomerization occurs at a photoactive double bond that defines the drive axis of the motor between the stator base and rotor head. Figure 1 shows the two isomers along with detail about the motor structure.

The photoisomerization of Z and E was studied in a solution of perdeuteroacetonitrile saturated with dry N₂ (to displace dissolved O₂ that can quench absorption). The solution is contained in a quartz 10 mm flat bottom NMR tube. The N₂ is introduced through a syringe that agitates the solution so that every molecule in the solution an equal opertunity to absorb a photon. The sample solutions initially contain only one isomer (Z or E) at a concentration of 5.0 x 10² M. This concentration is optimal for NMR analysis and for the photoisomerization experiments. Over the course of irradiation the single isomeric solution becomes a mixture of the two isomers due to photoisomerization. The reaction progress was measured using NMR. The absorption spectrum of the Z and E isomers is shown below in Figure 2.

The light source for the photoisomerization experiment was either a doubled dye laser pumped with the second harmonic of a 10 Hz Nd:YAG laser (for 280 and 320 nm measurements) or the direct fourth harmonic of the Nd:YAG laser (266 nm measurement). When using the dye laser a Pellin Broca prism is used to separate out all other wavelengths so that we have only the desired wavelength. The laser pulse widths were approximately 5 ns. The average pulse energy was measured during each irradiation cycle and was typically 2.0 mJ/pulse.

The laser light is aligned in such a way that the beam enters the NMR tube throught the flat bottom of the tube. All of the light is absorbed by the solution. The syringes are intoduced into the system through a septum attached to the top of the NMR tube. The longer syringe bubbles N₂ through the solution while the shorter syrings allows for out gassing of the N₂. The N₂ is saturated with perdeuteroacetonitrile by way of a prebubbler. The prebubbler is on ice to decrease the rate of evaporation of perdeuteroacetonitrile. Initial experiments used perdeuterodichloromethane, but results showed that this caused decomposition of our motor molecule most likely due to formation of hydorchloric acid. Figure 3 shows the experimental setup.

The near-UV spectrum has several peaks near 266, 280 and 320. Each of these were probed and the kinetic rates for the isomerization measured. Figure 4 is a typical set of NMR spectra. The small singlets located at 0.091 ppm and 1.39 ppm represent the Z and E isomers respectively. These peaks are monitored during the experiment to follow the photoisomerization, which shows a growth of the lesser isomer and decay of the greater isomer. This continues until a photostationary point is reached.

The vigorous agitation helps ensure that all molecules have an equal opportunity to absorb a photon over the course of an irradiation cycle (typically 0.25 - 1.0 min.). By knowing the laser puse energy, repetition rate, and the concentration of the solution, the average number of photons absorbed per unit irradiation time was obtained. The photoisomerization kinetics are reported below in terms of the average number of photons absorbed per molecule.

Experiments show that the molecular motor precursor photoisomerizes after absorbing a photon. Figure 5 shows the isomerization kinetics measured at several different wavelengths.

The experiment is repeated many times with the averages for the photoisomerization quantum yields Φ_(ZE) for isomer Z photoisomerizing to isomer E and Φ_(EZ) for the reverse process) tabulated below in Table 1. The ΔΦ values are the calculated error associated with the fits.

Experiments were also done using other 2-substitued molecules, an amino derivative and a pyrrol derivative. The NMR results for the pyrrol derivative are shown in Figure 6.

The photochemistry of the isomerization process is assumed to follow specific kinetic pathways. In the experiments shown there are two equations that correspond to the photodecomposition of the molecular motor precursor, both having quantum yields ~0. The photodecomposition was only present when perdeuterodichloromethane was used as a solvent. Data shown used perdeuteroacetonitrile which had neglegable decomposition in NMR measurements.

From these kinetic pathways we can calculate the Z and E concentrations over time assuming that their initial concentrations are [Z]₍₀₎ and [E]₍₀₎ for the respective isomers. The molar absorptivities of the Z and E isomers must also be including in the kinetics since they are unique to each molecule. If we assume that one of the isomeric concentrations is equal to zero prior to irradiation, measured to be zero using NMR, then we get the equations where as the number of photons per molecule approaches infinity, we reach the photostationary state. This gives equations which are ratios of the mole fractions. This ratio is scaled by the ratio of the molar absorptivities.

We can find a function from these equations that we can use to fit the experimental data. From the function we can calculate the photoisomerization quantum yields, Φ_(ZE) and Φ_(EZ). The quantum yields are actually the same as the rate constants if there is no decomposition and if we convert the units from `time' to 'number of photons absorbed'. The plots in Figure 5 illustrate the fit results for a specific set of data. The average results of this fitting method is tabulated in Table 1.

Synthesis of other motor precursors has been done to create derivatives of the 2-t-butyl-9-(2,2,2-tri-phenylethylidene)-fluorene compound by replacing the 2-t-butyl group with other substituents such as an amino, pyrrol, nitro, and phenanthrene groups. The hope is that these substituents will increase absorption feature wavelengths, thus allowing future experimental probing schemes to utilize the visible wavelength region. It would also be advantageous for these compounds to have a greater photoisomerization quantum yield so that the isomerization process would be more rapid.

The spectra of these molecules have been taken and are shown in Figure 7. The wavelength range has increased in the nitro compound into the low visible region possibly allowing the probing to be done with visible light.

Up to this point the kinetics for the 2-t-butyl, 2-amino, and 2-pyrrol substituted systems have been done. Unfortunately the amino and pyrrol systems do not show effective photoisomerization. This might be due to charge transfer mechanisms found in these new compounds that quench the photoisomerization activity before the motor can isomerize.

The experimental results show that for the 2-t-butyl-9-(2,2,2-tri-phenylethylidene)-fluorene does photoisomerize readily in solution. This is of great interest because the photo-activated drive unit may be very useful for molecular devices, including molecular motors. For any expectation of motor functionality this photoisomerization process must be verified prior to creating the chiral stator base. The molecules are thermally stable to isomerization up to the boiling point of the solvent (81.6^o) allowing us to be confident that the isomerization is due only to photoisomerization.. This is the first case of photoisomerization of a dibenzofulvene and, to our knowledge, of any fulvene or 1,1-diaryethene.

Experiments have also shown that the isomerization rates are wavelength dependent. Early analysis shows this may be due to the excitation of different electronic states of the molecule, therefore, resulting in different excited states promoting isomerization. This helps explain the great difference in the absorption spectrum for the 2-t-butyl substituted molecule at 266 nm and 320 nm.

The photoisomerization experiments tends to reach a photostationary point where the Z isomer is favored over the E isomer. The absorption spectrum shows that the E isomer has a greater molar absorptivity leading to the conclusion that the E isomer is a better absorber, hence, producing more of the Z isomer in the equilibrium. During synthesis the Z isomer is produced in a much higher yield prompting speculation that the Z isomer is a more stable molecule than is the E isomer. These two conclusions may work hand in hand to give us this isomeric bias.

The photoisomerization rate is also dependent upon the substituent placed at the 2 position on the rotor chromophore. Synthesis has produced several derivatives including an amino and pyrrol substituted molecule. It has been shown that the amino group can act as a quencher for fluorescence in biological molecules leading us to postulate that the amino substituted molecule is being quenched when in its excited state before isomerization can occur. Initial experiments also show that the pyrrol does not seem to isomerize either possibly due to the large bulk the pyrrol adds to the molecule causing an increase in the moment of inertia requiring more energy for photoisomerization. Steric hindrance may also be the culprit for why the pyrrol does not isomerize.