Abstract
We present the second order nonlinear optical properties of the Disperse Red 1 (DR1)-doped poly (cyano phenylene sulfide) (PCPS) - novel ferroelectric amorphous polymers. The PCPS possess self-organized long-range polarizations when they are annealed at temperatures higher than their glass transition point. From the unique nonelectrical poling effect, the second-order nonlinear susceptibility was obtained without the conventional poling procedures. The optimized conditions to create the second-order nonlinear optical susceptibility were also investigated in terms of the sample thicknesses and the annealing temperatures.
©2011 Optical Society of America
1. Introduction
Nonlinear optics (NLO) has been studied for almost half a century since the development of the laser. Nonlinear optical phenomena are defined as the simultaneous mixing of two or more photons inside the materials and the subsequent polarization formations with the sum of or the difference in frequencies of the photons participating in the interactions [1,2]. Many NLO active materials have been found or synthesized. The NLO active materials are roughly classified into three categories [3,4]. The first class, inorganic crystalline materials, are the most popular materials for practical usage, mainly because the fabrication techniques to produce large volume single crystals with high optical qualities are well established. The second class, organic crystalline materials [4], exhibit large nonlinear optical susceptibilities when the nonlinear optical interactions take place under the resonant conditions. The third class, polymer-based NLO materials, is inferior to the first two classes in terms of the NLO susceptibility and the thermal and mechanical properties. However, this class has several favorable properties such as the low weight, the low production cost and the ease of fabrication.
NLO polymers are typically composed of host and guest materials [4]. The guest materials are NLO chromophores with large molecular hyperpolarizabilities that participate in the nonlinear light - matter interactions. The host materials are amorphous polymers that sustain and fix the positions of the NLO chromophores. As initially prepared, the individual NLO chromophores are randomly oriented, and their structures are centro-symmetric. Hence, these materials exhibit the third-order NLO susceptibility but not the second-order.
The so-called “poling” procedure is performed to obtain the second-order NLO susceptibility in NLO polymers. This procedure typically involves applying an external field at temperatures that are higher than the glass transition point of the host materials to align the dipole moments of the NLO chromophores. However, dielectric breakdowns frequently occur during this procedure, because the electric fields as high as ~kV/mm are applied. Therefore, it is difficult to obtain the second-order NLO susceptibility in the polymer-based NLO materials with wide areas and large volumes.
Our previous study reported poly (cyano phenylene sulfide) (PCPS) as a novel host material for the NLO polymers (Fig. 1(a) ) [5]. PCPS exhibited several unique electrical behaviors [6]. First, it exhibited ferroelectric behaviors, which means the field - displacement hysteresis behaviors occur even in amorphous states. Second, the polarizations due to the long-range molecular dipole ordering were obtained simply through annealing at temperatures higher than the glass transition point. These phenomena may be referred to as “non-electrical poling” because the polar structures were formed without the external fields.
Because of the amorphous structure of the PCPS chains, NLO chromophores can be dispersed uniformly among the chains. Using the non-electrical poling techniques for the NLO chromophore-doped PCPS, the second-order NLO susceptibilities can likely be obtained without any applying external fields. Our previous study showed that the second-order NLO susceptibilities were successfully obtained from PCPS doped with Disperse Red 1 (DR1: N-Ethyl-N-(2-hydroxyethyl)-4-(4-nitrophenylazo)aniline), one of the most well-known NLO chromophores (Fig. 1(b)), simply by annealing at temperatures higher than the glass transition point of the PCPS. However, the previous study did not investigate the polarization formation dynamics of the NLO chromophores in the PCPS. The present work examines the second-order NLO susceptibility of the chromophores in terms of the film thickness and the annealing temperature.
2. Experimental
The PCPS was synthesized using a condensation polymerization method. The details were reported in our previous study [6]. The glass transition was observed at Tg~120 °C using a thermal analysis with differential scanning calorimetry. Disperse Red 1 (DR1) was chosen as the guest chromophore. For linear optical spectroscopy, the thin film samples were fabricated on the fused silica substrates with a spin coating method. The concentration of DR1 in PCPS was 10 w%. The samples used for NLO spectroscopy were also the thin films of the 10 w%-DR1-doped PCPS. The films were deposited on thin Au-coated fused silica substrates for nonelectrical poling; the surface energy difference between Au and air created an internal field for the polar orientations of the NLO chromophores. The film thicknesses were varied between 100 nm and 25 μm by controlling the rotation speeds of the spin coater and the concentrations of the solutions. The film thicknesses were determined by a surface profiler (Alpha-step IQ, KLA-tencor) with an accuracy of 1 nm. The Au-layers were prepared with vacuum deposition methods and had a typical thickness of 2 nm.
The linear absorption and the linear refractive index spectra of the samples were measured with a UV/VIS spectrometer (V-660, JASCO) and a spectroscopic ellipsometer (FZVWC, Mizojiri Optical Co. Ltd.), respectively. The refractive index was determined by comparing with the quantity of the fused silica substrate, which was established in the previous study [7]. The analysis was conducted, assuming the fixed thickness of the polymer layers.
The second-order NLO susceptibilities were studied using the second harmonic (SH) methods. Optical pulses from a Ti: Sapphire regenerative amplifier were used as the excitation beams. The center wavelength, pulse energy, pulse width and repetition rate were 800 nm, 0.5 mJ, 50 fs and 1 kHz, respectively. The polarizations of the excitation beams were controlled with a λ/2 wave plate. The SH signals were separated from the residual excitation beams by color glass filters (B460, HOYA) and then detected with a photomultiplier tube (1P28, Hamamatsu Photonics K.K.). The polarization of the SH signal was resolved using a Gran-Taylor prism, and the p-polarized component was measured.
3. Results and discussion
Figure 2(a) shows the linear absorption spectra of the DR1-doped PCPS thin films with a 2.0 μm-thickness. The linear absorption spectrum of the pure PCPS thin film with a 200 nm-thickness is also shown as a reference. The undoped PCPS film had a peak at 340 nm due to its cyano-benzene moieties. In contrast, the DR1-doped PCPS thin films exhibited an absorption peak at 520 nm due to the π-π* transitions in the Disperse Red 1 [8,9]. When the SH methods were employed, the wavelengths of the fundamental excitation and the frequency-doubled beams were 800 and 400 nm, respectively. The DR1-doped thin film was almost transparent at 800 nm, and had a weak absorption at 400 nm. A part of the SH signal emitted from DR1 was reabsorbed. The absorption spectrum of the DR1-doped PCPS thin films was not well resolved at wavelengths shorter than 350 nm, because of the intense absorption band due to the PCPS. Figure 2(b) shows the spectrum of the linear refractive index of the DR1-doped PCPS films with a thickness of 300 nm.
The SH signal intensities from the DR1-doped PCPS thin films were observed continuously during the annealing procedures. The samples were decolorized when they were annealed at temperatures above 170 °C, presumably due to the decomposition or the sublimation of the DR1 chromophores. Hence, the measurements were performed at temperatures lower than 150 °C. The samples were heated from 40 to 150 °C and then cooled from 150 to 40 °C. Both the heating and the cooling processes were performed at a rate of 3 °C/min. The incidence angle of the excitation beam was 45 °. The excitation beam was p-polarized, and the intensities of the p-polarized SH signal were recorded.
Figure 3 shows the SH signal intensity () of the DR1-doped PCPS thin films as a function of temperature. The film thickness was 2.3 μm. The samples did not emit significant SH signal before annealing. During heating, the SH signal appeared at approximately 120 °C and increased monotonously. Because the glass transition temperature of the PCPS was ~120 °C, the polarization formations of the NLO chromophores were likely associated with the rearrangement of the chain segments above the glass transition. During the subsequent cooling step, the intensity of the SH signals decreased slightly but remained constant below 120 °C. The same measurement was performed for the non-doped PCPS films with the thicknesses between 100 nm and 20 μm, but no SH signal was recorded. Hence, it was concluded that the SH signals from the DR1-doped PCPS were due to the second order NLO susceptibilities of the DR1 in polar order.
Weak SH signals were emitted from some of the samples prior to annealing but these signals were not reproducible. It was tentatively concluded that this second order NLO susceptibility was due to a short-range dipole ordering caused by the flow orientations during the evaporation of the solvents. For the purposes of this study, this phenomenon was not investigated further.
To determine the degree of polar order of the NLO chromophores, the SH signal intensity was measured with varying polarization angles and the p-polarized SH signals were recorded. Figure 4 shows the measured values for the sample with a 2.3-μm film thickness. The signal intensity was greatest at = 0° and lowest at = 90 °. The trace was reproduced with a model based on the C∞v symmetry of the rod-like NLO chromophores, which were non-centro-symmetrically oriented in the poled thin film [10].
In this model, the macroscopic second order NLO susceptibility is predominantly determined by the tensor component of the chromosphere molecular hyperpolarizability where the ζ-axis denotes the molecular long axis. The dipole moments of the NLO chromophores are oriented along the direction of the external field but are slightly tilted against the normal line of the substrate surface. The NLO molecules are arranged along the normal line of the substrate plane surface and are distributed randomly within the substrate plane surface. The macroscopic second order NLO susceptibility results from the polar order of the NLO chromophores along the normal line.
In general, the second-order NLO susceptibility has 27 tensor components. Among them, there are only three non-zero independent components in the present model with a C∞v symmetry. Setting the z-axis as the normal line of the substrate surface and the x- and y-axes as the two axes within the substrate surface, these non-zero components are expressed as , , and ( or ). These components can be expressed as a function of the tensor component and the molecular tilt angle , which is measured with respect to the substrate surface normal (Fig. 5(a) ).
Here, is the density of the NLO chromophores and is a local field factor in the direction at frequency . The molecular tilt angle is related to the first-order parameter for the polar orientations of the NLO chromophores as [11,12].When the fundamental beams that propagate through the x-z plane with a -polarization angle excite the samples at the incidence angle (Fig. 5(b)), the intensity of the p-polarized SH signals is expressed by Eq. (2).where and are the refractive index of the polymer and the refracted angle at frequency , respectively. The coefficients , , , and in Eq. (3) are related to the Fresnel refractions between air and the polymer layer.Because the film thicknesses of the samples prepared in the present study are close to the wavelength of the fundamental beam or wider, the term is included to account for the phase-mismatch between the fundamental beam and the frequency-doubled beam in the polymer.Assuming that the Kleinman’s condition is satisfied, is approximately equal to [13]. In the present case, this condition may not be valid, because of the weak absorption component at 400 nm, which is the wavelength of the SH light. However, many previous studies successfully determined the second-order NLO susceptibilities under the Kleinman’s condition even in the near-resonant regions, similar to the present study [14,15]. Thus, the analysis was conducted under the assumptions of these conditions. Under this approximation, the molecular tilt angle is related to the second order NLO susceptibility tensor components as
Thus, the molecular tilt angle can be determined after obtaining the ratio by fitting the trace to Eq. (2). For the sample with the thickness = 2.3 μm in Fig. 4, the tilt angle was determined to be = 19.8°. The NLO chromophores were well oriented along the direction normal to the substrate plane.The trace was recorded for samples with several different film thicknesses. For these measurements, the samples were annealed at 150 °C, and the pump beams were incident at = 45°. The SH intensities at = 0 and molecular tilt angles are plotted as functions of the film thickness in Fig. 6 . As mentioned above, there is a weak absorption at 400 nm, the wavelength of the SH light, and the SH signal was attenuated by the reabsorptions. Here, the influence of the reabsorption was corrected. The quantity , which is proportional to the SH conversion efficiency, has a peak near = 10 μm and decreases with the thickness. The minimum molecular tilt angle occurred at = 5 μm.
The same measurement was conducted for a z-cut quartz plates to determine the nonlinear optical coefficient , which corresponds to . Here, is dielectric constant. By comparing with the SH signal of the z-cut quartz that was measured with the same optical geometry, the coefficient was determined to be ~0.6 pm/V for the sample with = 2 μm thickness. The quantity was by two magnitude smaller than = 27 pm/V from the poled polymer that consisted of the DR1-grafted onto the poly (methyl methacylate) [16]. The results indicate that not all of the NLO chromophores took part in the polarization formations. A significant portion of the NLO chromophores probably oriented such that the dipole moments were anti-parallel to one another.
In this study, the second-order NLO susceptibilities were obtained through annealing at temperatures higher than the glass transition temperature. Furthermore, the NLO chromophores were highly oriented along the normal of the substrate plane surfaces in the annealed polymer samples. These data indicate that the second-order NLO susceptibility is due to the non-electrical poling effect proposed in our previous study [6].
The segmental motions of the PCPS were allowed, and they redistributed their conformations at temperatures higher than Tg. An internal field was applied from the thin film interface with the Au-layer, which has a higher surface energy, to the air interface, which has a lower surface energy. At temperatures higher than , the PCPS chains changed their conformations to align the molecular dipole moments of their cyano groups with the internal field. The molecular dipole moments of the DR1 molecules, surrounded by PCPS chains, also likely aligned their orientations along the macroscopic polarization of the PCPS cyano dipoles. Hence, the polar orientations of the DR1 resulted in the formation of a macroscopic second order nonlinear optical susceptibility. Once the polar conformations of PCPS and DR1 were formed, these conformations were frozen during the cooling step. Consequently, the macroscopic second order NLO susceptibility remained even after cooling.
However, the nonelectrical poling model cannot consistently explain the film thickness dependence of the molecular tilt angle or the polar order parameters of the NLO chromophores shown in Fig. 6. The molecular tilt angle was not monotonically dependent on the film thickness; a minimum of the tilt angle was observed at = 5 μm. However, the nonelectrical poling model predicts that the intensity of the internal field should be proportional to the difference in the surface energy and the inverse of the film thickness. Hence, as the film thickness became thinner and the intensity of the internal field became higher, the order parameter of the NLO chromophores should have risen, or the molecular tilt angle should have decreased.
These unexpectedly wider molecular tilt angles in the thinner films were likely due to the surface roughness of the Au layers on the substrate. Atomic force microscopy showed that the surface morphology of the Au layer was uneven and consisted of aggregations of small spherical Au particles with an average diameter of 20 nm. Chains of PCPS adsorbed on the uneven surfaces of the Au layer were prevented from being packed neatly along the normal line of the substrate plane surface. Therefore, the PCPS chains and the NLO chromophores adopted less ordered conformations near the Au layer. As the distance from the surface increased and the influence of surface roughness lessened, the chains were able to pack more neatly and adopt more ordered conformations. As a result, the overall order parameter increased for the thicker samples.
The SH conversion efficiency dependence on the film thickness was seemingly reproduced with the term in Eq. (4), which represents the influence of phase-mismatching. The highest conversion efficiency was obtained at . The condition was satisfied when the film thickness was equal to the phase-matching coherence length defined as
Using the refractive index = 1.493 for the fundamental frequency and = 1.672 for the SH frequency in Fig. 2(b), the coherence length was determined to be = 1.0 μm. However, the highest SH conversion efficiency was recorded from the samples with the film thickness = 10 μm (Fig. 6(a)), ten times longer than.The origin of the inconsistency between the two quantities has not been clearly identified. The inhomogeneous polarization distribution within the film thickness is one of the most probable reasons for the inconsistency, as in the case of poly (vinylidene fluoride) thin films [17,18]; the films were only partly polarized, and the second-order NLO interactions occurred in only limited parts of the thin films. In this case, the apparent phase matching coherence lengths of the inhomogeneously poled thin samples would be longer than those of the homogeneously poled ones. As discussed above, the polarization formations of the PCPS chains likely occurred first in the Au layer and propagated toward the air-exposed layer. The present results may indicate that the polarization of the NLO chromophores did not grow within the entire film, and it was limited to the vicinity of the Au interface.
4. Conclusions
We have observed the second-order NLO properties in non-electrically-poled DR1-doped PCPS films. These results show that the non-electrical poling techniques are suitable to obtain the second-order NLO susceptibilities in films with μm thicknesses simply by annealing at temperatures higher than the glass transition temperature without an external electric field. The polar order of the NLO chromophores was determined by the balance between the internal field strength and the polymer-metal interfacial interactions. The optimal SH conversion efficiency was obtained with a film thickness ten times wider than the phase-matching coherence lengths. This may be because the films were only partly polarized and the second harmonic susceptibilities arose from limited portions of the films. Hence, the polar order of the NLO chromophores seemed to be relatively short-ranged in the non-electrically poled polymer in the present study, which is different from the case in the electrically poled polymers. In spite of this disadvantage, the non-electrical poling techniques and their ability to facilitate the second-order NLO susceptibility within the polymers will be useful for further developments of polymer optical sciences and technologies.
Acknowledgments
This work was supported by the Murata Science Foundations and the Japan Society for Promotion of Science.
References and links
1. Y. R. Schen, The Principles of Nonlinear Optics (John Wiley & Sons, 2003).
2. R. W. Boyd, Nonlinear Optics, 3rd ed. (Academic Press, 2008).
3. V. G. Dmitriev, G. G. Gurzadyan, and D. N. Nikogosyan, Handbook of Nonlinear Optical Crystals (Springer-Verlag, 1998).
4. H. S. Nalwa, T. Watanabe, and S. Miyata, “Organic materials for nonlinear optics,” in Nonlinear Optics of Organic Molecules and Polymers, H. S. Nalwa and S. Miyata, eds. (CRC Press, 1996), pp. 89–350.
5. A. Sugita, M. Morimoto, Y. Ishida, N. Mase, Y. Kawata, and S. Tasaka, “Linear and nonlinear optical properties of disperse red dyes in poly-(cyano phenylene sulfide),” Chem. Phys. Lett. 501(1–3), 39–43 (2010). [CrossRef]
6. J. Ide, S. Tasaka, and N. Inagaki, “Nonelectrical poling in ferroelectric polycyanophenylenesulfides,” Jpn. J. Appl. Phys. 38(Part 1, No. 4A), 2049–2052 (1999). [CrossRef]
7. I. H. Malitson, “Interspecimen comparison of the refractive index of fused silica,” J. Opt. Soc. Am. 55(10), 1205–1208 (1965). [CrossRef]
8. L. De Boni, C. Toro, A. E. Masunov, and F. E. Hernández, “Untangling the excited states of DR1 in solution: an experimental and theoretical study,” J. Phys. Chem. A 112(17), 3886–3890 (2008). [CrossRef] [PubMed]
9. C. Toro, A. Thibert, L. De Boni, A. E. Masunov, and F. E. Hernández, “Fluorescence emission of disperse Red 1 in solution at room temperature,” J. Phys. Chem. B 112(3), 929–937 (2008). [CrossRef] [PubMed]
10. K. Kajikawa, H. Takezoe, and A. Fukuda, “Symmetry and second order susceptibility of hemicyanine monolayer studied by surface second harmonic generations,” Jpn. J. Appl. Phys. 30(Part 1, No. 5), 1050–1062 (1991). [CrossRef]
11. F. Lagugné-Labarthet, C. Sourisseau, R. D. Schaller, R. J. Saykally, and P. Rochon, “Chromophore orientations in a nonlinear optical azopolymer diffraction grating: even and odd order parameters from far-field Raman and near-field second harmonic generation microscopies,” J. Phys. Chem. B 108(44), 17059–17068 (2004). [CrossRef]
12. Y. Tu, Q. Zhang, and H. Ågren, “Electric field poled polymeric nonlinear optical systems: molecular dynamics simulations of poly(methyl methacrylate) doped with disperse red chromophores,” J. Phys. Chem. B 111(14), 3591–3598 (2007). [CrossRef] [PubMed]
13. D. A. Kleinman, “Nonlinear dielectric polarization in optical media,” Phys. Rev. 126(6), 1977–1979 (1962). [CrossRef]
14. T. G. Zhang, C. H. Zhang, and G. K. Wong, “Determination of molecular orientation in molecular monolayers by second-harmonic generation,” J. Opt. Soc. Am. B 7(6), 902–907 (1990). [CrossRef]
15. K. Shirota, K. Kajikawa, H. Takezoe, and A. Fukuda, “Molecular orientation in mixed layers of hemicyanine and fatty acid at air/water interface studied by second harmonic generations,” Jpn. J. Appl. Phys. 29(Part 1, No. 4), 750–755 (1990). [CrossRef]
16. T. Pliška, W.-R. Cho, J. Meier, A.-C. Le Duff, V. Ricci, A. Otomo, M. Canva, G. I. Stegeman, P. Raimond, and F. Kajzar, “Comparative study of nonlinear-optical polymers for guided-wave second-harmonic generation at telecommunication wavelengths,” J. Opt. Soc. Am. B 17(9), 1554–1564 (2000). [CrossRef]
17. E. Bihler, K. Holdik, and W. Eisenmenger, “Polarization distributions in isotropic, stretched or annealed PVDF films,” IEEE Trans. Electr. Insul. 24(3), 541–545 (1989). [CrossRef]
18. S. Bauer, G. Eberle, W. Eisenmenger, and H. Schlaich, “Second-harmonic generation with partially poled polymers,” Opt. Lett. 18(1), 16–18 (1993). [CrossRef] [PubMed]