532

J. Opt. Soc. Am. B / Vol. 24, No. 3 / March 2007

Gindre et al.

Toward submicrometer optical storage through controlled molecular disorder in azo-dye copolymer films Denis Gindre,* Alex Boeglin, Gregory Taupier, Olivier Crégut, Jean-Pierre Vola, Alberto Barsella, Loïc Mager, Alain Fort, and Kokou D. Dorkenoo Institut de Physique et Chimie des Matériaux de Strasbourg, UMR 7504 Université Louis Pasteur-CNRS, 23 rue du Lœss, Boîte Postale 43, F-67034 Strasbourg CEDEX 2, France Received July 18, 2006; revised September 22, 2006; accepted October 25, 2006; posted November 9, 2006 (Doc. ID 73125); published February 15, 2007 While information encoding through optically induced orientation of azo dyes in organic matrices is being extensively pursued, we propose the alternative of starting out with poled films and to locally reduce their second-harmonic generation capacity by a focused near-infrared femtosecond laser beam of moderate intensity. Arrays of dots irradiated under varying conditions are subsequently imaged in situ as dark spots on a bright background. The samples are also examined through conventional optical microscopy and through atomic force microscopy. We demonstrate that, of these techniques, second-harmonic imaging performs best in the task of information retrieval. © 2007 Optical Society of America OCIS codes: 190.4160, 190.4180, 190.4710, 210.4810.

1. INTRODUCTION A key feature of organic materials is their plasticity, not only at a macroscopic scale but all the way down to the molecular level. For instance, it is possible to induce local changes in the refractive index of a solid cross-linked polymer matrix by increasing the density of the medium through additional polymerization. If the latter is triggered by two-photon absorption (TPA) from a tightly focused near-infrared laser beam, submicrometer voxels can be etched into the volume of a 3D sample1,2 to create structures like waveguides.3 An entirely different way to alter optical properties in organic media is to induce dichroism or birefringence through light induced structural and orientational changes among chromophore units embedded in various polymers, gels, or resins.4,5 In this respect, the two-photon isomerization of disperse red 1 (DR1) guest molecules, far below the glass transition temperature of a solid film of poly(methyl methacrylate) (PMMA) where spontaneous large amplitude motions are otherwise frozen, together with the spatial resolution inherent to TPA processes, has recently been proposed to record information in the form of orientational hole burning in dots that can subsequently be read through confocal differential reflection microscopy.6 The association of DR1 to PMMA, either as a host– guest system or in the grafted form of a copolymer, has been investigated intensively for well over a decade with the promise of applications in photonics (optical switching,7 electro-optic modulator,8 microactuators,9 surface gratings,10 etc.). Because of its relatively good stability, it has become a model system in quadratic nonlinear optics (NLO) of polymeric materials where it is essential to achieve at least a partial noncentrosymmetric orientational order to get electro-optic activity and second0740-3224/07/030532-6/$15.00

harmonic generation (SHG) capability. In the 1990s, Fiorini et al.11 extensively studied the all-optical poling of DR1 grafted on PMMA copolymer films where the chromophore undergoes photoisomerization cycles that are driven by resonant one- and two-photon absorption processes. The coherent superposition of the fundamental frequency of an IR laser source with the second harmonic generated in a phase-matched type-II crystal produces a quasi-permanent polar alignment that is spatially modulated with a period satisfying the phase-matching condition for SHG. A further refinement of this technique consists in the control of the angle of linear polarization for each frequency component of the irradiating fields that allows the tailoring of the symmetry of the resulting quadratic tensor ␹2. This scheme has been exploited recently by Bidault et al.12 to demonstrate that information can be stored in areas differing in multipolar polarization pattern which can later be discriminated, i.e., read, through the detection of SHG intensity while scanning the sample area with the linearly polarized IR beam alone. Each of the above techniques to encode information into an organic polymer film through the optical orientation of its azo-dye component undoubtedly has its merits. However, it appears that there may well be a much simpler way to achieve this goal: provided that the polymer film has been previously corona poled13 to become, in essence, a blank or white page, a focused IR beam may be used to locally decrease the degree of alignment of the chromophores. Since the writing step now consists in depoling a small volume within the material, a standard multiphoton microscopy setup can pick up the SHG signal, which can be used to monitor the loss of ␹2, either to study the dynamics of the phenomenon, or to serve as a feedback to control the writing process. © 2007 Optical Society of America

Gindre et al.

Vol. 24, No. 3 / March 2007 / J. Opt. Soc. Am. B

533

2. EXPERIMENTAL To explore the performance of this controlled erasing scheme, we have chosen to work with samples prepared from a commercially purchased (Specific Polymer, molecular weight= 20,000 g / mol, Tg = 125° C) methyl– methacrylate copolymer where one monomer out of ten carries a DR1 molecule. In practice, a 20 wt. % solution of DR1-functionalized PMMA in 1,1,2-trichloroethan is spin coated on a transparent microscope slide and left to dry at 80° C for 1 h yielding films with a typical thickness of 400 nm as measured by a DekTak Profilometer. The film is then corona poled13 at elevated temperature on a heating plate at 85° C in a normal atmosphere: a voltage of 4.7 kV is applied between the grounded aluminum plate below the glass slide and two tungsten wires above the polymeric film while the sample is allowed to slowly cool down to room temperature over a period of about 2 h. This procedure yields 2 ⫻ 2 cm2 samples generating a SHG signal that is nearly uniform. Both the SHG imaging and the TPA-induced isomerization (local depoling) are performed at room temperature using an unamplified, linearly polarized, and tunable 共670– 1100 nm兲 Ti:sapphire laser source (120 fs pulse duration, 80 MHz repetition rate). A half-wave plate and a polarizer are used to adjust the intensity of the beam before it is focused onto the sample by a microscope objective 共NA= 0.45兲. Larger apertures may be used to achieve a better lateral resolution, but a spot diameter of about 1 ␮m for the illuminated area is sufficient for our purpose. Under these conditions, the thickness of our films, namely 400 nm, is smaller than depth of field (Rayleigh length). The SHG signal is collected by another microscope objective before being filtered by a dichroic beam splitter and sent to a spectrograph where a liquid nitrogen cooled CCD camera records its spectrum. The sensitivity of the present setup is such that 800 nm laser pulses of typically 0.1 nJ in energy (corresponding to an average pulse power of 1 kW or a maximal intensity of about 15 GW/ cm2 at the center of the focal point) will yield a SHG signal recorded by the camera as a 10 nm FWHM spectral band centered at ⬃400 nm presenting a signal-to-noise ratio of 10. This is more than adequate to verify through a complete 2D scan of the film that the spin-coating and corona-poling processes have produced a uniform sample.

3. DEPOLING EFFICIENCIES If, instead of performing the scan mentioned above, we keep the sample fixed for any length of time, the camera records the evolution of the SHG efficiency of an area ⬃1 ␮m across. Such real time SHG efficiency decay curves are plotted in Fig. 1(a) for several values of illumination power. When attempting to fit the results with a Debye relaxation function, it becomes quickly obvious that the observed time behavior is strongly nonexponential. Of course, this is to be expected when following dynamical phenomena in samples made of an amorphous polymeric material. Indeed, most experimental measurements performed in glassy materials, be they spectroscopic, dielectric, or viscoelastic in nature are routinely

Fig. 1. (Color online) Real time SHG efficiency decay and relaxation rate of the second-order nonlinear susceptibility. (a) Normalized SHG intensity as a function of time for six values of the illumination power from 1.0 to 5.6 mW. (b) Relaxation rate of the second-order nonlinear susceptibility versus illumination power. Experimental data (black squares, also shown in log/log representation in the inset) are extracted from depoling curves with a stretched exponential relaxation function. Curves in (b) correspond to a quadratic law.

interpreted with the help of a small number of empirical nonexponential mathematical laws.14 Of prominent use for experiments carried out in the time domain is the Kohlraush–Williams–Watts “stretched exponential” relaxation function15,16 that introduces two parameters, a stretching exponent 共␤兲 and a relaxation time 共␶兲:

冋 冉 冊册

冑I2␻共t兲 = 冑I2␻共0兲 exp

t

−

␶

␤

,

0 ⬍ ␤ 艋 1.

共1兲

When used to fit the collected data, we observe only a small dispersion among the exponent ␤ = 0.50± 0.05. It is therefore sensible to extract the relaxation rates, i.e., the inverse of the relaxation times ␶ [Fig. 1(b)]. These rates agree to better than 10% with a quadratic power law with respect to the input power in near-IR laser pulses. We are thus lead to identify TPA within the DR1 chromophores themselves, followed by isomerization along the azo axis, as the main driving force behind the loss in SHG signal. Furthermore, the quadratic dependence of the loss rates with respect to pulse energy dismisses heating through residual one-photon absorption in the sample as a detectable contribution to the depoling process. In this, we are in agreement with the recent orientational hole-burning experiments6 under comparable illumination conditions but on different samples, namely, 5 ␮m thick films of a

534

J. Opt. Soc. Am. B / Vol. 24, No. 3 / March 2007

10% DR1 in PMMA guest–host system instead of the thin copolymer films investigated here. Instead of following the dynamics of local SHG efficiency loss over long time spans, an improved characterization method consists in analyzing the effect of writing operations performed for finite time durations. Figure 2(a) shows the real time recording of the SHG signal accompanying a series of focused irradiations carried out for

Gindre et al.

a fixed time span of 300 ms at 785 nm. While the shutter is closed, the sample is translated by 6.7 ␮m and the halfwave plate in front of the polarizer is rotated to modify the input power for the following stationary burn in. Illumination power was first reduced from 10 mW down to 0.3 mW then raised up again to 10 mW in 109 settings for the angle of the half-wave plate. This procedure has been adopted in order to concentrate the low-intensity points in

Fig. 2. (Color online) Evolution of SHG signal versus illumination power. (a) Real time recording of the SHG signal during the burn-in of 109 dots: 300 ms exposure time, power settings between 0.3 and 10 mW. (b) and (c) Zoom-in on the data points for powers ⬃1.8 and 8.7 mW, respectively. (d) SHG intensities at the opening (䊏) and before the closing (쎲) of the shutter [see arrows in (b) and (c)] versus illumination power. (e) SHG contrast for illumination powers over 10 mW displays a saturation effect.

Gindre et al.

the center of the strip to reduce possible alignment artifacts. Enlargements of this recording [Figs. 2(b) and 2(c)] demonstrate that at low intensities, no net SHG signal loss is detectable during the 300 ms exposition time, whereas at higher illumination powers, the decay dynamics become apparent even though the integration time at the CCD camera is 40 ms. The data points [Fig. 2(d)] reflect the behavior of the SHG signal as a function of pulse energy at low intensities. In other words, with this technique we have direct evidence that the phenomenon is indeed quadratic in laser intensity, while benefiting in addition from better sensitivity at lower intensities where fitting the SHG decay curves over long times becomes problematic. At high pulse power, both methods to assess the dynamics will ultimately fail because of the finite response time of the camera. To circumvent this obstacle, we proceed now in two stages. We start by writing a row of dots using a fixed illumination time of 300 ms at the same wavelength as above while increasing the pulse energy, from 10 to 65 mW. Once the sample had been written on in this manner, we conduct a 10 ␮m / s scan of the strip with the same laser beam at 2.1 mW to record the SHG signal curve. During this reading step, every point of the strip is therefore exposed a second time to 785 nm light for about 100 ms (assuming a 1 ␮m spot diameter). Thus this step promotes further depoling in the sample, and therefore induces its own signal. Nevertheless, when extracting the maximum and minimum SHG signal values from a 5 ␮m stretch over each written spot, one gets the rather clean normalized SHG contrast as a function of laser intensity [Fig. 2(e)]. For data storage applications, it is this contrast that will be the relevant parameter to determine the proper writing and reading conditions. A sharp increase in contrast is followed by the onset of saturation consistent with the progressive exhaustion in chromophores that have not been reoriented during the initial writing step.

4. TWO-DIMENSIONAL IMAGING ANALYSIS To judge the performance of our controlled depoling scheme as a means of optical storage, it is desirable to investigate the spatial response of the material by 2D imaging techniques. Standard optical microscopy has been used to detect changes in refractive index in the thin films while atomic force microscopy (AFM) has been performed to assess changes in the surface morphology of the samples. Some results are displayed (Fig. 3) where pictures from a Nanoscope AFM set are shown to the left, an image from an optical microscope is in the center and our SHG data is to the right. This particular sample has been marked with 20 lines of 14 dots separated by 7.5 ␮m (except between the leftmost two rows, which are only spaced by half this distance). In addition to a square pattern, the writing protocol called for a gradual increase in incident power from 1.3 to 48.1 mW but was otherwise the same as the one previously stated. The first thing one should notice in Fig. 3 is that the thresholds for detecting written spots with AFM and optical microscopy are almost identical. The AFM results show the presence of pits

Vol. 24, No. 3 / March 2007 / J. Opt. Soc. Am. B

535

Fig. 3. (Color online) Optical and AFM images of printed dots. (a) Optical microscopy of an array of 20 lines of dots written with illumination power increasing from 1.3 (top left) to 48.1 mW (bottom right). (b) SHG imaging scan of the same area of the sample with a power of 2 mW. (c)–(e) AFM scans of three areas depicted by white squares in hole depths of (c) 4, (d) 7, and (e) 20 nm. Small black specks (typically 200 nm in diameter) appear during film preparation and poling. Scale bars in all pictures are 7.5 ␮m.

on the surface of the sample with typical depths of 4 [Fig. 3(c)], 7 [Fig. 3(d)], and 20 nm [Fig. 3(e)]. This decrease is small compared with the 400 nm thickness of the film and is primarily ascribed to the relaxation of metastable spatial configurations in the polymer. The resulting increase in density of the material accounts for the change in refractive index detected in optical microscopy. SHG imaging, on the other hand, is not sensitive to this densification per se, but only to the loss in acentric order among the chromophores. Whether these two phenomena should be thought of as inextricably linked or somewhat independent of each other may be a matter of opinion, but the fact remains that the detection of the SHG signal is the more sensitive technique. This is clearly observed by inspecting the upper parts of images [Figs. 3(a) and 3(b)]. Thus it is possible to store optical information in a small, controlled amount of local disorder that can only be detected through SHG imaging. The two-photon depoling and subsequent SHG imaging turns out to be surprisingly immune to problems with the quality of the spin-coated thin films. Figure 4(a) shows the SHG image of a blank (poled) sample of rather dubious quality. This area has then been marked with an array of dots using the experimental settings given in the caption. The upper row [Figs. 4(b)–4(e)] shows four subsequent scans performed with an increasing spatial resolution. Figure 4 also illustrates the high quality of 3D representations obtained from a more uniform sample. Using a representation scheme that appears to turn the holes inside out is a way to further reduce variations in background SHG intensity as the fake color representation demonstrates [Fig. 4(f)]. The reader is referred to the caption for the exact settings used. Of course, other patterns, such as the 50 ␮m wide grating strip with a 4 ␮m step [Fig. 4(g)], can be burned into the material as well. In this instance, a dark frame had been burned around the grating itself using a higher incident power of over 100 mW. Such permanent markings are useful when erasing the written patterns by reheating and repoling the samples anew since otherwise it would be difficult to find the same area again.

536

J. Opt. Soc. Am. B / Vol. 24, No. 3 / March 2007

Gindre et al.

Fig. 4. (Color online) SHG pictures of several patterns. (a) 200⫻ 200 ␮m area of a blank poled sample. (b)–(e) Four successive scans after burn-in of 10 ␮m spaced dots with ␭ = 800 nm at an illumination power of 50 mW, exposure time of 200 ms. All scans performed at 4.2 mW with steps of (b) 4, (c) 1, (d) 0.5, and (e) 0.4 ␮m. (f) Inverted representation of SHG loss imaging in fake colors (writing power of 50 mW and reading power of 3.6 mW). (g) Grating with a 4 ␮m step (etching at 10.3 mW and scanning at 3.1 mW).

5. CONCLUSIONS The possibility to induce a controlled amount of disorder within a limited volume of corona-poled thin polymeric films containing polar chromophores is of practical interest because such a domain can subsequently be imaged with great sensitivity thanks to the contrast in SHG signal with respect to its immediate surroundings. Since the local depoling is achieved using focused near-infrared femtosecond laser pulses to trigger the isomerization of an azo dye through TPA processes, submicrometer resolution is readily achieved, and the same holds true for the SHG imaging technique since it is also quadratic in field intensity. The experiments were carried out at wavelengths near 800 nm where the TPA cross section of the PMMA–DR1 copolymer exhibits a minimum so that, although we operate at a high repetition rate and are limited in CCD integration time and shutter speed, it is possible to keep the deposited energy low enough to prevent irreversible damages. Provided that we avoid large changes in surface morphology, the written spots can be erased by repoling above the glass transition temperature. Indeed, it has been possible to burn dots and image them a second time around and, although the contrast suffered in the process, two-photon depoling of azo-dye copolymers holds the promise for an erasable data storage device. Of course, we also take advantage of the low TPA cross section of DR1 ⬃800 nm when performing the SHG imaging scans where transparency at half that wavelength may seem an issue. However, this problem can be sidestepped by reading the remaining SHG efficiency with laser pulses of a different frequency, outside the twophoton absorption band of the chromophore. Obviously, the proper tailoring of spectral properties of azo dyes through chemical syntheses is also called for in this respect. An additional opportunity lies in the association of

two different chromophores: an azo dye whose photoisomerization cycles will affect the orientation of another NLO molecule through induced motion in the PMMA backbone. In short, two-photon depoling offers flexibility in assembling composite materials, while SHG microscopy provides an efficient diagnostics tool to rate their performance.

ACKNOWLEDGMENTS This work was supported by the Region Alsace and by Centre National de la Recherche Scientifique (CNRS). K. D. Dorkenoo is the corresponding author and can be reached via e-mail at [email protected].

*Present address, Laboratoire des Propriétés Optiques des Matériaux et Applications, UMR CNRS 6136 Université d’Angers, 2 boulevard Lavoisier, F-49045 Angers CEDEX.

REFERENCES 1.

2. 3.

4.

W. Zhou, S. M. Kuebler, K. L. Braun, T. Yu, J. K. Cammack, C. K. Ober, J. W. Perry, and S. R. Marder, “An efficient two-photon-generated photoacid applied to positive-tone 3D microfabrication,” Science 296, 1106–1109 (2002). S. Kawata, H. B. Sun, T. Tanaka, and K. Takada, “Finer features for functional microdevices,” Nature 412, 697–698 (2001). S. Klein, A. Barsella, H. Leblond, H. Bulou, A. Fort, C. Andraud, G. Lemercier, J.-C. Mulatier, and K. D. Dorkenoo, “One-step waveguide and optical circuit writing in photopolymerizable materials processed by two-photon absorption,” Appl. Phys. Lett. 86, 2111181–2111183 (2005). N. Bohm, A. Materny, H. Steins, M. M. Muller, and G. Schottner, “Optically induced dichroism and birefringence

Gindre et al.

5. 6.

7. 8.

9.

10.

of disperse red 1 in hybrid polymers,” Macromolecules 31, 4265–4271 (1998). C. Fiorini, F. Charra, A. C. Etilé, P. Raimond, and J. M. Nunzi, Advanced Functional Molecules and Polymers, H. S. Nalwa, ed. (Gordon & Breach, 2001), pp. 167–216. M. Maeda, H. Ishitobi, Z. Sekkat, and S. Kawata, “Polarization storage by nonlinear orientational hole burning in azo dye-containing polymer films,” Appl. Phys. Lett. 85, 351–353 (2004). Y. Luo, W. She, S. Wu, F. Zeng, and S. Yao, “Improvement of all-optical switching effect based on azobenzene-containing polymer films,” Appl. Phys. B 80, 77–80 (2005). A. Leinse, M. B. J. Diemeer, A. Rousseau, and A. Driessen, “A novel high-speed polymeric EO modulator based on a combination of a microring resonator and an MZI,” IEEE Photon. Technol. Lett. 17, 2074–2076 (2005). Y. L. Yu and T. Ikeda, “Alignment modulation of azobenzene-containing liquid crystal systems by photochemical reactions,” J. Photochem. Photobiol. C 5, 247–265 (2004). P. Rochon, E. Batalla, and A. Natansohn, “Optically induced surface gratings on azoaromatic polymer films,” Appl. Phys. Lett. 66, 136–138 (1995).

Vol. 24, No. 3 / March 2007 / J. Opt. Soc. Am. B 11.

12.

13.

14.

15. 16.

537

C. Fiorini, F. Charra, J. M. Nunzi, and P. Raimond, “Quasipermanent all-optical encoding of noncentrosymmetry in azo-dye polymers,” J. Opt. Soc. Am. B 14, 1984–2003 (1997). S. Bidault, J. Gouya, S. Brasselet, and J. Zyss, “Encoding multipolar polarization patterns by optical poling in polymers: towards nonlinear optical memories,” Opt. Express 13, 505–510 (2005). M. A. Mortazavi, A. Knoesen, S. T. Kowel, B. G. Higgins, and A. Dienes, “Second-harmonic generation and absorption studies of polymer-dye films oriented by coronaonset poling at elevated temperatures,” J. Opt. Soc. Am. B 6, 733–741 (1989). R. Hilfer, “H-function representations for stretched exponential relaxation and non-Debye susceptibilities in glassy systems,” Phys. Rev. E 65, 0615101–0615105 (2002). G. Williams and D. C. Watts, “Non-symmetrical dielectric relaxation behaviour arising from a simple empirical decay function,” Trans. Faraday Soc. 66, 80–85 (1970). G. T. Boyd, “Optical second-harmonic generation as an orientational probe in poled polymers,” Thin Solid Films 152, 295–304 (1987).