Open Access
Add to BibSonomy
Abstract
In order to investigate the mechanism of photo-induced processes of azo molecules doped in polymer, temporal evolution of the absorption spectrum and transmittance for two polarization components were measured under and after optical pumping. A simple model was applied to decompose the contributions from angular hole burning (AHB) due to photo-isomerization and following molecular reorientation (MRO). The result for disperse red 1 (DR1) in poly-methyl methacrylate (PMMA) indicated the dominance of AHB in the order of seconds while slow accumulation of the MRO effect became significant during a longer time range. Several extraordinary behaviors observed in preceding studies will be explained in context given in this study. Dependence on dye concentration showed that MRO was more significant in highly loaded samples. Further, quantum yield of photo-isomerization process in this system was estimated to be 8%.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Azo-doped and azo-functionalized polymers are promising for optical data processing, mass storage and dynamic holography. In these thirty years, extensive experimental and theoretical results have accumulated, attracting interest of scientists in many areas as optoelectronics, mechanical and microfabrication engineering, because photo-stimulated molecular motion and macroscopic deformation made it possible to form unique nanoscale structures applicable to novel devices as photonic crystals, metamaterials and so on [1,2].
It has been widely believed that optically induced birefringence and dichroism in azo-containing films were caused by trans-cis photo-isomerization and following molecular reorientation occurring in relaxation from cis to trans forms. Thus, repetition of excitation-relaxation made the molecular orientation statistically perpendicular to the excitation light polarization leading to strong anisotropy. Furthermore, continuous excitation induced fluidity of the materials resulting in macroscopic deformation of the polymer surfaces which have been observed as surface relief gratings in many azo systems [3,4].
Since these processes are so complex that many phenomenological parameters are required to describe their temporal evolution, complete understanding of the mechanism has not been established. The most popular techniques employed in basic studies were photo-induced birefringence (PIB), four wave mixing (FWM), and self-diffraction observed in degenerate two-wave mixing (TWM) geometry, because these phenomena are tightly related to practical application and also precise measurements were possible due to their background free configurations. However, some drawbacks inherent in these methods make it difficult to interpret experimental results. For PIB, anisotropy of refractive index is caused by the population change of the two isomers, and also by the redistribution of molecular direction. And matrix motion also induces PIB showing complex behaviors in rising and relaxation processes [5–7]. Furthermore, PIB reflects the difference between two index components, so it is impossible to determine the sign of change in each component. FWM and self-diffraction also suffer from similar problems. In these cases, signal intensity is determined by the index change in illuminated region and it is impossible to know the sense of it. In addition, both real and imaginary parts of the optical constant contribute, making the analysis more difficult especially when the materials have light absorption at probe wavelength [8–11].
In order to overcome the drawbacks, several efforts were made to separate the refractive index and absorption effects by direct measurement of each constant before and after optical excitation. In very early days, Sekkat and Dumont obtained the change of orthogonal components of refractive index for Disperse Red 1 (DR1) doped in poly-methyl methacrylate (PMMA) with attenuated total reflection technique [12]. In their study, the polymer was excited by a linearly polarized light, and refractive indices were measured by a probe laser with parallel and perpendicular light filed. Result indicated the reduction of the index for parallel direction and also small reduction in the other direction. These changes cannot be understood by simple models incorporating only angular hole burning (AHB) or molecular reorientation (MRO) as described later. On the other hand, polarization dependence in absorbance change and its temporal evolution was studied by a couple of groups. Todorov et al. measured absorption for methyl orange doped in a polymer under optical excitation, showing decrease and increase for parallel and perpendicular components, respectively. The result supported the quick molecular alignment to the direction perpendicular to the polarization [13]. Contrary to that, Rodrigues et al. monitored probe light transmittance for DR1 polymer under optical excitation finding strong reduction in parallel direction and weak reduction in perpendicular direction [14]. Blanche et al. observed complicated behaviors in absorbance change for several azo polymers at multiple wavelengths claiming the importance of glass transition temperature of polymer and the absorption in cis state [15].
In this study, we made direct measurement of absorbance evolution during and after optical excitation in DR1 doped PMMA thin films. Because the absorption measurement was relatively simple and it only reflects the dynamics of dopant, this method is one of the most adequate ways to get direct insight into dye behaviors. In the next section, sample preparation and experimental setups are described. Results on spectrum evolution and precise traces of absorption change at the excitation wavelength are given as well as PIB signals in section 3. In section 4, based on a simple model associated to AHB and MRO, contributions from the two mechanisms are discriminated showing that AHB with decay time of 1sec was dominant in dilutely doped systems and the effect from MRO became significant if the excitation continued for several minutes in highly loaded samples.
2. Experimental
2.1 Sample preparation
DR1 and PMMA (Mw:120,000) were purchased from Sigma-Aldrich and used without further purification. Molecular structure of the dye is shown in Fig. 1. Among fabricated films with various concentrations, preparation procedure for a 2wt% case is explained as an example. The dye and polymer were dissolved in chloroform with concentration of 2 and 100g per liter, respectively. Films were formed from the solution with spin coating of 2,000rpm on a glass substrate (25 x 25mm), giving clear films having 2wt% dye and 0.95μm thickness, the value was obtained with a surface profiler DEKTAK-6M. Samples were dried in a vacuum chamber for 3 hours at 60°C. Absorption spectrum of the film measured with a spectrometer (Shimadzu UV-1800) is shown in Fig. 1 along with two other concentration cases. Since absorbance at 532nm was 0.1492, nearly homogeneous excitation along the depth direction was possible. For comparison, samples with 1, 5, 10 and 20wt% dye were prepared by adjusting the amount of solvent and spin speed to make absorbance values at 532nm be in the range between 0.1~0.2 as given in Table 1.
Fig. 1 Molecular structure of DR1 and absorption spectra for PMMA films containing 2, 5, and 20wt% of the dye.
Table 1. DR1/PMMA samples prepared from chloroform solutions.
2.2 Experimental setup
Instrument for in situ absorption measurement is shown in Fig. 2(a), where a halogen lamp was used as a source and a fiber coupled linear-sensor-equipped spectrometer (USB2000, Ocean Optics) was employed for quick measurement. Minimum temporal resolution for the system was about ms, but typically hundred times averaging limited actual resolution to about 1sec. In our experimental condition, measurement was made every 5sec. Absorbance charts were obtained from the spectra of the transmitted light divided by a light source spectrum measured with a glass substrate only.
Fig. 2 Experimental setups for (a) in situ polarized absorption spectrum measurement under and after external excitation, and for (b) simultaneous measurements of absorbance depletion and induced birefringence under optical excitation. S: mechanical shutter, D: detector. All setups were placed in a dark room.
For excitation, vertically polarized 2nd harmonics from a cw YAG laser (Lumine, GLM) was used. Its intensity was distributed in nearly Gaussian shape with 0.325mm radius for half intensity, and peak intensity was estimated to be 105mW/cm2. White probe light was polarized by a conventional film polarizer and its intensity was the order of μW which was weak enough not to disturb the dye states. An electromagnetic shutter was controlled manually. Measurement was made during laser exposure of 1min and about 10min after shutter closing.
The second optical setup in Fig. 2(b) was made to observe absorbance change and birefringence induced by cw excitation in a faster time scale. As shown in the figure, four beams overlapped on sample surface. A YAG 2nd harmonics (Coherent, DPSS532) was used for excitation, and other two probes with orthogonal polarizations were generated from separate YAG sources (Lumine, GLM) in order to avoid the interference among them. The film was set between two cross-polarized Glan-Thompson prisms to detect PIB with a He-Ne laser which was linearly polarized with 45° to the polar axis of the pump. Absorbance at 633nm for the samples was negligible. Photodiodes (PD) were used to monitor the intensities of four beams. Outputs were led to a digital oscilloscope DL1620 (Yokogawa) with 1MΩ input impedance and a termination resistor was mounted when necessary. In front of every PD head, an iris with beam size aperture was placed to eliminate contamination of light as far as possible. Temporal resolution was determined by the shutter speed 0.5ms.
3. Results
3.1 Induced change in the absorption spectrum
Absorption spectra for 2wt% film (sample 2) at several instances before, during and after excitation are picked up to show in Fig. 3(a) in which case the probe polarization was parallel to the pump. Absorbance reduced strongly by the pump almost instantaneously and recovered quickly after blocking the excitation. Spectrum approached back close to its initial shape in 10min after excitation, but there remained small deviation. On the other hand, when polarization was perpendicular, reduction was about one third of that for the parallel case, and the spectrum almost completely recovered original magnitude after 10min as depicted in Fig. 3(b).
Fig. 3 Temporal evolution of absorption spectrum for DR1(2wt%)/PMMA before, during and after excitation with linearly polarized pump beam. Probe light was polarized (a) parallel and (b) perpendicular to pump. Dips observed at around 532nm were due to scattering of excitation light. Time dependences of absorbance at peak wavelength for (c) DR1(2wt%)/PMMA and (d) DR1(20%)/PMMA.
Absorption peak intensities are plotted against time in Fig. 3(c), clearly showing the difference between two polarization cases. For the parallel direction, slowly developing component was observed during excitation while the perpendicular counterpart had no such slow growth. In decay process, there seemed to be at least three components as sudden depletion, exponential-looking decay with time constant of several 10sec, and very long lifetime component which was missing in the perpendicular case. However, the long life component was not persistent, since the spectrum completely recovered after left in dark for one day. When dye concentration was higher, the changes in perpendicular direction became small relative to the parallel counterparts as shown in Fig. 3(d) for the case of 20%. In the higher loaded samples, inclination to the opposite direction during irradiation became significant for the perpendicular case. By changing the dye concentration, gradual transition from Fig. 3(c) to (d) was observed in medium concentration cases.
3.2 Time dependence of polarized absorbance and birefringence
Output voltages obtained with D1 and D4 in Fig. 2(b) gave extinction coefficients and that from D2 did birefringence. If the output from the detectors is expressed by Vi(t) at an instance t from excitation onset, extinction coefficient κ is calculated by the following expression.
Here, λ and d indicate the wavelength and film thickness, respectively, and α10(0) means absorbance of the sample obtained before the measurement. If phase retardation due to birefringence is much less than unity, the difference of two refractive indices can be approximated by Eq. (2) using the output from D2.
where V0 is the calibrated voltage corresponding to the light intensity obtained when direction of the analyzer in Fig. 2(b) is parallel to the polarizer before pump irradiation. Typical signals are shown in Fig. 4(a) for the sample 2, which was irradiated for 60sec repeatedly after 60sec breaks with a pump intensity of 35mW/cm2. The behaviors of κ in two polarizations shown in Fig. 4(b) were consistent to those in Fig. 3 where long lifetime in parallel direction and relatively quick response in perpendicular component were observed. It is reasonable that the birefringence shape reflected to the larger component of κ, since it is given by the subtraction of two components even though these are the real parts. The results and those on other samples were analyzed with a method given in section 4, and concentration dependence will be also discussed.Fig. 4 Temporal evolution of (a) PIB and (b) extinction coefficients parallel and perpendicular to excitation polarization in DR1(2wt%)/PMMA film.
3.3 Long time behavior in absorbance
In order to study the evolution during long time excitation, continuous monitoring of transmission light was made for 30min. Figure 5 shows the change of extinction coefficient κ for the sample 2 under the exposure to the pumping beam with peak intensity of 135mW/cm2. It seemed that the process was composed of fast and slow components approaching to a constant value as long as the linear scaled data Fig. 5(a) was shown. However, the same data plotted with log-scaled time axis indicated very slow evolution which has been continuing even at 30min after the irradiation starting as shown in Fig. 5(b).
Fig. 5 Temporal evolution of extinction coefficient for DR1(2wt%)/PMMA thin film under excitation with 135mW/cm2 pump beam of 532nm, plotted with (a) linear and (b) logarithmic temporal scales.
4. Discussion
To understand the results described in the previous section, we considered that the changes of κ was caused only by the combination of AHB and MRO. And DR1 molecule was assumed to be a thin rod with one definite dipole moment along the molecular axis. We also presumed that the absorption by cis molecules was negligible at 532nm.
At first, let us consider the case where AHB occurs without reorientation. When randomly oriented molecules are excited by linearly polarized light, probability of photon absorption is proportional to cos2θ, where θ is the angle between molecular axis and optical electric field. Consequently, parallel oriented molecules are selectively excited and parts of them relax to cis state reducing the absorbance and refractive index in parallel direction. At the same time, perpendicular component also reduces since every molecule can be excited due to its parallel directional component unless it aligns completely perpendicular. Total magnitude of the absorbance reduction in perpendicular direction can be calculated statistically when saturation is negligible, giving one-third of that for the parallel counterpart. The other extreme limit is the case where reorientation is dominant, and it is the situation when cis molecules quickly recover trans form with change of their direction. In this process, we assume that reduced amount of absorbance in parallel direction is equally distributed to two other axes [16,17].
Providing that total change of κ is the summation of these two mechanisms, it is possible to retrieve two components from the absorbance observed in two polar directions with the following equations.
Here, indicates the change of κ from the value given at t = 0.Equation (3) is the same but a factor to the reduction of total absorbance which has been used to estimate the annihilation of trans isomers [17,18]. Instead of simple subtraction which has been used as a degree of dichroism so far, factor 3 was multiplied to the perpendicular term to retrieve MRO effect. These equations converted the data in Fig. 4(b) into those in Fig. 6(a) which showed quick rise and decay in the AHB component without accumulation effects in spite of intermittently continuing excitation for 15min. Its relaxation process shown in Fig. 6(b) indicates that the decay constant was about 1sec which was corresponding to the value having been assigned to the lifetime of cis state [5,19]. Several longer lifetime components were also observed, which might have been assigned to long lived cis state or recovery of re-oriented molecules in several former studies [7,20]. We conclude that it reflected the recovery process from cis state which was delayed due to steric hindrance or spatial restriction.
Fig. 6 (a) Contribution from AHB and MRO for change of extinction coefficient on DR1(2wt%)/PMMA. The data was retrieved from those in Fig. 4(b). (b) shows decay process after the last excitation period. Lines are drawn to show two components with the shortest time constants.
The component assigned to MRO in Fig. 6(a) showed increase during excitation and decrease in break periods, and the curve also gave abrupt changes at the instances of shutter operation including a sudden decrease at the beginning of the first irradiation period. If orientation change is caused by molecular motion in cis state or at the instance of backward transition to trans state, change at these instances would be moderate all along the process. Although there might be physical effects triggered by excitation on-off, the most possible explanation would be an artifact due to the deviation from the assumptions. In the derivation of Eqs. (3) and (4), we neglected saturation effect which would be significant if Δκ/κ was not very small. Indeed, when large depletion of absorbance was observed due to relatively strong pumping, the jumps became more significant. However, it is possible to estimate the contribution from MRO through time averaging or envelopes of the curves.
From this analysis, one can say that the seemingly complete recovery observed in Fig. 3(c) and 4(b) for perpendicular direction after blocking excitation was an accidental cancelation of AHB and MRO components. Extraordinary behaviors observed in the preceding studies might have been explained in the same context [12,14].
We made the comparison among the systems with different dye concentrations. Rough estimation for Δκ due to AHB gave values of 0.0025, 0.016, 0.017 and 0.009 for 2, 5, 10, and 20% samples, respectively. The data for the last one is depicted in Fig. 7, from which the value 0.009 was extracted from the bottom and top envelops of the black curve. MRO components were 0.001, 0.011, 0.018, and 0.013, respectively. Although the values were not proportional to the dye concentration, the contribution from MRO was known to be more significant for more concentrated specimens. Similarly the maximum of birefringence was given as 0.0013, 0.013, 0.013, and 0.015 for the same series. The results suggested that intermolecular interaction played a role for molecular orientation, as the slight peak shift observed in absorption spectra in Fig. 1 indicated the existence of intermolecular interaction. However, since molecular motion also correlates to the type of polymer and its molecular weight, conclusion should be withheld until more detailed studies are made focusing on matrix effects.
As discussed, temporal behavior of extinction coefficient was rather complex even after the analysis, so it seems to be difficult to make complete description in the whole time range. However it might be possible to claim important characteristics by paying attention on certain time regions. Because all molecules lay in trans state before light irradiation, the change in very early stage would be determined only by the initial depopulation of trans state molecules. By extrapolating the initial decay shown in Fig. 5(a), the efficiency or quantum yield of photo-isomerization upon photon absorption can be estimated from initial slopes.
A method to derive the quantum yield of photo-isomerization has been established for the case of dyes in solutions where induced anisotropy cancels instantaneously by thermal motion. Loucif-Saibi et al. have estimated the quantum yield of DR1 in PMMA to be 10% by applying the method using optical power dependence of cis and trans population ratio in equilibrium under strong pumping [21,22]. Our estimation by using absorbance depletion was rather simple, although we assumed that the absorbance of cis state was negligible at the exciting wavelength. Instead, we took into account the polarization dependence of absorbance providing the molecule was characterized by a simple dipole. Thus, we assumed that collision cross section of the molecule with a linearly polarized photon was approximated by σcos2θ, where θ meant the angle between dipole axis and the polarization direction and σ had the dimension of area. When molecules distribute randomly in orientation, absorbance is expressed as
Here, λ is wavelength and N0 is the number of molecules in a unit volume. Photo-isomerization selectively reduces the number of molecules having small θ, and absorbance depletion depends on probe polarization which can be calculated statistically [16,17]. When the probe beam has the same polar axis to the pump, absorption coefficient is expressed by the following equation when only the zeroth and first order terms are taken into account.
Np means incident photon number per unit area and Φ is the quantum yield of photo-isomerization reaction. Number of incident photons until extrapolation line crosses the time axis (0.8s in the present case of 2wt% sample, see Fig. 5(a)) can be calculated from light intensity. Employing the cross section value estimated from absorbance before excitation and the dye concentration for the samples, Φ was finally estimated to be 8.6% for 2wt% sample and 8.2% for 1wt% sample. Considering the uncertainties in film thickness, density of the polymer and the accuracy of extrapolation, these two values are well coincident. If we did not consider the orientation dependence on collision cross section as justified in solutions, estimated value would be 9/5 times larger than these.
5. Conclusion
Optically induced birefringence and dichroism in DR1 doped in PMMA thin films were investigated by absorption and pump-probe measurements under and after optical excitation with polarized light, showing depletion in absorbance in both polar components. With the simple model, the contributions from AHB and MRO were separated properly, confirming quick rise and decay due to photo-isomerization and relatively slow growth and slow decay by MRO process. In the more concentrated dye system, the more significant the contribution from MRO was. Long term non-persistent motion was found which could not be expressed by superposition of exponential processes. Quantum efficiency of photo-isomerization was estimated to be 8%. These experimental and analysis methods will be useful for the studies on the basic characteristics of photo-induced processes in azo polymers to improve their efficiency as optically functional devices. Since great simplification was made to formulate the analysis, further sophistication of the model is a next issue.
Funding
JSPS KAKENHI (JP17K06033).
Acknowledgments
A part of measurement was conducted with equipment supported by Nanotechnology Platform Program (Synthesis of Molecules and Materials) of MEXT, Japan.
References and links
1. Y. Zhao and T. Ikeda, eds., Smart Light-responsive Materials (Wiley, 2009).
2. S. Lee, J. Shin, Y.-H. Lee, S. Fan, and J.-K. Park, “Directional photofluidization lithography for nanoarchitectures with controlled shapes and sizes,” Nano Lett. 10(1), 296–304 (2010). [CrossRef] [PubMed]
3. P. Rochon, E. Batalla, and A. Natansohn, “Optically induced surface gratings on azoaromatic polymer films,” Appl. Phys. Lett. 66(2), 136–138 (1995). [CrossRef]
4. D. Y. Kim, S. K. Tripathy, L. Li, and J. Kumar, “Laser-induced holographic surface relief gratings on nonlinear optical polymer films,” Appl. Phys. Lett. 66(10), 1166–1168 (1995). [CrossRef]
5. P. Rochon, J. Gosselin, A. Natansohn, and S. Xie, “Optically induced and erased birefringence and dichroism in azoaromatic polymers,” Appl. Phys. Lett. 60(1), 4–5 (1992). [CrossRef]
6. A. Priimagi, M. Kaivola, F. J. Rodriguez, and M. Kauranen, “Enhanced photoinduced birefringence in polymer-dye complexes: hydrogen bonding makes a difference,” Appl. Phys. Lett. 90(12), 121103 (2007). [CrossRef]
7. H. Ono, N. Kowatari, and N. Kawatsuki, “Study on dynamics of laser-induced birefringence in azo dye doped polymer films,” Opt. Mater. 15(1), 33–39 (2000). [CrossRef]
8. C. A. Walsh and W. E. Moerner, “Two-beam coupling measurements of grating phase in a photorefractive polymer,” J. Opt. Soc. Am. B 9(9), 1642–1647 (1992). [CrossRef]
9. D. Sek, E. Schab-Balcerzak, M. Solyga, and A. Miniewicz, “Polarisation-sensitive holographic recording in polyimide-containing azo-dye,” Synth. Met. 127(1–3), 89–93 (2002). [CrossRef]
10. A. Sobolewska and S. Barkiewicz, “Three gratings coupling during the holographic grating recording process in azobenzene-functionalized polymer,” Appl. Phys. Lett. 92(25), 253305 (2008). [CrossRef]
11. Y. Kawabe, K. Fukuzawa, T. Uemura, K. Matsuura, T. Yoshikawa, J. Nishide, and H. Sasabe, “Photoinduced grating formation in a polymer containing azo-carbazole dyes,” Appl. Opt. 51(27), 6653–6660 (2012). [CrossRef] [PubMed]
12. Z. Sekkat and M. Dumont, “Polarization effects in photoisomerization of azo dyes in polymeric films,” Appl. Phys. B 53(2), 121–123 (1991). [CrossRef]
13. T. Todorov, L. Nikolova, and N. Tomova, “Polarization holography. 1: a new high-efficiency organic material with reversible photoinduced birefringence,” Appl. Opt. 23(23), 4309–4312 (1984). [CrossRef] [PubMed]
14. A. Rodriguez, G. Vitrant, P. A. Chollet, and F. Kajzar, “Photochromic properties of azo dye polymer thin films: evidence of an additional reversible mechanism of molecular behavior,” Rev. Mex. Fis. 52(2), 135–138 (2006).
15. P.-A. Blanche, P. C. Lemaire, M. Dumont, and M. Fischer, “Photoinduced orientation of azo dye in various polymer matrices,” Opt. Lett. 24(19), 1349–1351 (1999). [CrossRef] [PubMed]
16. Z. Sekkat and M. Dumont, “Photoinduced orientation of azo dyes in polymeric films. Characterization of molecular angular mobility,” Synth. Met. 54(1–3), 373–381 (1993). [CrossRef]
17. M. Dumont, S. Hosotte, G. Froc, and Z. Sekkat, “Orientational manipulation of chromophores through photoisomerization,” Proc. SPIE 2042, 2–13 (1994). [CrossRef]
18. M. Dumont and A. El Osman, “On spontaneous and photoinduced orientational mobility of dye molecules in polymers,” Chem. Phys. 245(1–3), 437–462 (1999). [CrossRef]
19. Z. Sekkat, D. Morichere, M. Dumont, R. Loucif-Saibi, and J. A. Delaire, “Photoisomerization of azobenzene derivatives in polymeric thin films,” J. Appl. Phys. 71(3), 1543–1545 (1992). [CrossRef]
20. R. Hildebrandt, M. Heigelich, H.-M. Keller, G. Marowsky, S. Hvilsted, N. C. R. Holme, and P. S. Ramanujam, “Time-resolved investigation of photoinduced birefringence in azobenzene side-chain polymer films,” Phys. Rev. Lett. 81(25), 5548–5551 (1998). [CrossRef]
21. R. Loucif-Saibi, K. Nakatani, J. A. Delaire, M. Dumont, and Z. Sekkat, “Photoisomerization and second harmonic generation in disperse red one-doped and -functionalized poly(methyl methacrylate) films,” Chem. Mater. 5(2), 229–236 (1993). [CrossRef]
22. H. Rau, G. Greiner, G. Gauglitz, and H. Meier, “Photochemical quantum yields in the A (+h.nu.). dblarw. B (+h.nu.,DELTA.) system when only the spectrum of A is known,” J. Phys. Chem. 94(17), 6523–6524 (1990). [CrossRef]
