We report the direct high-voltage poling of the chromophores 4-methoxy-4’-nitrostilbene (MNS) and 2-methyl-4-nitroaniline (MNA) in a 1.2 mm thick polymethylmethacrylate (PMMA) host. The DC fields used to pole the guest-host system varied in strength from 58 to 67 Vμm−1, and the presence of chromophore orientation in the poled samples was subsequently analyzed via Maker fringe analysis. We observed values of the SHG coefficient d 33 of 0.7 to 1.75pm/V at 532nm.
© 2011 OSA
Poled polymers have been proposed for use in a number of electro-optics applications, such as telecommunications switching devices and devices that both emit and detect terahertz (THz) radiation. During the poling process, the chromophores in a guest-host polymer system spontaneously align in a strong applied electric field. The poling process is usually carried out near the glass transition temperature (Tg) of the polymer host, i.e., the point at which the polymer host softens, allowing chromophore rotation and alignment to occur. However, when cooled below Tg of the polymer host the aligned chromophores become locked in their new orientation within the polymer host. When poled in this way, guest-host polymers can exhibit many useful electro-optic properties. One application where a longer path length of poled guest-host polymer clearly results in enhanced performance of the electro-optic device is in THz signal generation and detection. Many researchers have demonstrated that a longer optical path length, which is available in bulk poled polymers, allows improved THz power, sensitivity and signal-to-noise ratio in both THz generators and detectors [1–5]. Large-area Pockels cells constructed from electro-optic guest-host polymers will find new uses in high-speed modulators and shutters. In addition, with greatly improved materials, bulk-poled guest-host polymers may ultimately replace inorganic nonlinear optical (NLO) crystals in certain applications, such as second-harmonic generation (SHG), optical parametric amplification/generation (OPA/OPG) applications and frequency mixing. To this end, we report on the thickest known section of guest-host polymer poled by any technique. At 1.2 mm, this thickness is nearly an order of magnitude greater than the thickest samples poled for THz generation and detection.
An alternative approach to the high-voltage direct-poling technique described in this paper is optical poling [6–10]. This method has also been shown to accomplish spontaneous chromophore alignment in thick sections of a guest-host polymer. Rather than relying on a DC electric field as described in this work, optical poling relies on the large average electric field available in some CW lasers and in femtosecond pulses. Unfortunately, optical poling has only been demonstrated in chromophores that undergo an optically-induced cis-transisomerization such as Disperse Red 1 (DR1) and cannot be used for chromophores such as those described in this paper, or those used in poled guest-host polymers for THz generation and detection.
Faced with the limitations of optical poling and the difficulties of processing and stacking large numbers of extremely thin corona-poled films for bulk applications, we chose the more direct approach of high-voltage DC contact poling . This method may, in principle, be employed to pole any polar chromophore in a guest-host system. Indeed, the current work demonstrates poling of chromophores in samples nearly twice as thick as those described in . To be sure, the larger electrode separation distance in slabs exceeding 1 mm thickness necessitates the use of much higher poling voltages than in thin films. These high voltages are potentially lethal and require extreme caution together with the use of a high-voltage power supply with current overlimit protection.
We chose the chromophores used in this work both for their ready availability in quantity, either from commercial sources or via easy synthesis, as well as for their relative transparency in the visible region of the spectrum. The latter criterion reflects our interest in all-polymer electro-optic devices with visible-spectrum applications. In general, chromophores with a transmission window in the visible possess lower μβ values than those with a transmission window in the infrared, on poling yield polymers with lower bulk EO coefficients, and have been the subject of considerably less study.
2.1 Chromophore and guest-host polymer sample preparation
The chromophores used in this study were 4-methoxy-4ˊ-nitrostilbene (MNS) and 2-methyl-4-nitroaniline (MNA), shown in Fig. 1 . The MNA was obtained from Sigma Aldrich and recrystallized from methanol prior to compounding. Pelletized polymethylmethacrylate (PMMA) was produced by CYRO Industries (Acrylite H15) and distributed by AMCO Plastics. All other chemicals were obtained from Sigma Aldrich or Alfa Aesar and were used as received without further purification. MNS was synthesized according to procedures in references [12,13].
The production of guest-host polymer samples above 1 mm in thickness is greatly facilitated by a compounder and an injection molding machine; conventional thin-film polymer techniques such as spin casting or solution casting are not optimal. Spin casting can deposit homogeneous films but builds thickness relatively slowly; conventional casting techniques frequently resulted in films with aggregated chromophore as the solvent evaporated. To facilitate production of these samples, an approximately 1:1 mixture of PMMA and chromophore (wt/wt) was dissolved in a common solvent (acetone) and cast into an aluminum tray (45 mm diameter). As the solvent evaporated, tack-free films formed over a period of about 1 day at room temperature in a fume hood and were further dried in an oven at 60° C to drive off residual solvent. The films were broken into small pieces and fed into the compounder to achieve the desired chromophore content as described below.
The polymer samples were processed on a DSM Xplore 15 cc Micro-Compounder, consisting of a clamshell barrel with two conical screws and a 15 mL recirculation pathway to allow for extended processing intervals. The Xplore system is equipped with a pressure transducer which measures the downforce exerted by the polymer melt, roughly correlating with the viscosity of the melt. Sample plaques were prepared using a DSM Xplore 12 mL Injection Molding Machine, equipped with a mold yielding 2.5 cm squares, approximately 1.3 mm thick.
A representative procedure for the preparation of a compounded guest/host system isdescribed below. Beginning with a pre-heated (235° C) clean barrel and screws, a known quantity of virgin PMMA was used to purge the compounder and the recirculation pathway, removing any residual particulates or polymer from the cleaning process. The extrudate was recovered and weighed, yielding a measure of remaining polymer in the compounder (5.260g). The pre-mixed chromophore/polymer blend and virgin PMMA were charged into the compounder (1.554 g chromophore, 1.604 g PMMA from pre-mixed blend, and 6.623 g virgin PMMA) to provide approximately 15 g (15.041 g, 10.3% wt/wt chromophore) total material for processing, and screw rotation was maintained at 100 rpm. After the final addition of the chromophore and PMMA, the feed-port was closed and the polymer melt was compounded for 15 min and maintained under a gentle N2 purge at about 5 L/min. During this time, the barrel temperature was reduced from 235° C to 200° C because of plasticization of the PMMA by the chromophore. The viscosity of the melt dropped significantly after chromophore addition, and a drop in temperature was required to permit sufficient shearing of the melt to ensure homogenization of the blend. After processing for 15 min, the extrudate was collected in the heated transfer line/plunger assembly used by the DSM Xplore 12 mL injection molding system. The transfer line had been pre-heated to 230° C, and the extrudate was fed into the transfer line until the pressure transducer measured less than 600 N. The transfer line was then placed into the cradle for the injection molder and the polymer was injected into a room temperature mold using an 8 bar/10 bar injection pressure (8 bar initial impulse with 10 bar following to ensure full mold fill). Total injection cycle time for each sample was about 10 sec. The parts were removed from the mold immediately by applying pressure to the runner of the injected part, and no mold release was required. Individual plaques were removed from the injection tab and were then wrapped in aluminum foil to prevent damage to the surface of the plaque. A small sample was broken off from the injection runner to permit analysis by DSC for Tg determination. The Tg values were measured using a Perkin Elmer Q1000 DSC, with a 0° C - 150° C heat-cool-heat cycle, heating at 10° C/min heat and cooling at 20° C/min. The transition temperature is reported as the inflection point of the transition taken from the second heat cycle. Measured values of Tg were 75.3° C, 83.0° C, and 76.9° C for the 10% MNA, 10% MNS, and 15% MNS, respectively.
2.2 Poling of bulk guest-host polymer samples
Chromophore alignment was accomplished by the use of a poling fixture of our own invention (Fig. 2 ), which consists of two nylon blocks fabricated into a clamp-like device, secured by two nylon screws. Polished copper electrodes, each 2 cm in diameter, were inserted into the fixture through holes drilled into the clamp. The copper electrodes have rounded edges to prevent undesired electric field concentrations. The copper electrodes were carefully polished as flat as possible; during the poling process the smooth surfaces were impressed upon the polymer plaques, compressing the 1.3 mm thick plaque down to about 1.2 mm, eliminating the need for any post-poling polishing steps. The two Nylon screws allow one to adjust the tension against the polymer plaque and to ensure that the copper electrodes are seated parallel to the faces of the plaque. High-voltage connections were made by securing the high-voltage leads from the high-voltage power supply to the poling fixture by copper alligator clamps. This was done so that the poling fixture can easily be removed from the silicone oil bath to access the sample. The copper electrodes are smaller (2 cm) than the polymer plaques (2.5 cm) to allow the extra polymer material to increase the hold-off voltage of the polymer in the clamp.
A Spellman 120 kV high-voltage DC power supply (Model SL120P60) was used to supply the high voltage required for poling. The power supply has an adjustable current overlimit setting that is designed to prevent continuous arcing. The overlimit was set at about 250 μA. Once the limit is exceeded, an internal circuit breaker shuts off the high voltage. The poling fixture was connected to the power supply as described above and then immersed in Dow-Corning 710 silicone oil contained in a 2L Pyrex reaction flask. A Teflon cup was placed in the bath to prevent arcing to the walls of the oven, and Teflon sheathing lines the inside of the reaction flask. The entire bath/sample assembly was then placed in a Cascade Tek TO-3 oven for poling. The temperature was controlled with a Watlow temperature controller. The temperature was set to the inflection point of the glass transition temperature Tg, with the temperature rising from ambient to Tg over the course of two hours and then maintained near Tg for approximately 10 minutes. The voltage was increased to 70 to 90 kV and applied throughout the entire poling process. As the electric fields near the clamp are large and any thermocouple device will act as an undesirable ground and suffer damage, the temperature next to the poling clamp was measured with a conventional mercury thermometer. The current through the plaques ranged from 1 to 5 μA during poling, with the plaques conducting higher currents as the temperature approached Tg. At the end of the poling process, when the temperature reached approximately 40° C, the samples were removed from the oil bath and trimmed of excess unpoled material.
2.3 Maker fringe measurement and analysis
The Maker fringe experiment shown in Fig. 3 employs a variant of the optical setup described by Maker et al. . A 1064 nm EKSPLA laser operating at 10 Hz and a 30 psec pulse width was focused to a spot 40 μm in radius (half-width 1/e2) on the sample using a 750 mm lens. Energy at the sample was 30-35 μJ, well below the polymer damage threshold of approximately 100 μJ for the laser spot size used. A co-aligned 633 nm HeNe laser was introduced into the beam path through the back of one of the 1064 nm dielectric mirrors, and an iris was used to establish normal incidence and search for a spot on the sample with low surface scatter. The iris was also used to establish co-linearity between the 633 nm and 1064 nm beams. The first half-wave plate was used to adjust power on the reference meter. By using a polarizing beamsplitter cube as a pickoff, the existing degree of polarization in the transmitted beam is retained or enhanced. P-polarized light was obtained by minimizing reflection at Brewster's angle through the adjustment of the second half wave plate. The second waveplate was used to establish the p-polarized state of the incident laser in the sample. Two KG glass filters provided >10 OD of attenuation at 1064 nm with around 90% transmittance at 532 nm, attenuating the transmitted pump light to well below the 0.1 pJ threshold of the Laser Probe RJP-765 silicon energy meters. The energy meter, with a 1 cm diameter active area, was placed 16 cm from the sample to gather as much of the scattered light as possible. The pump energy at 1064 nm was measured throughout the experiment by a calibrated reference meter. Maker fringe data for an x-cut quartz crystal of approximately 0.5 mm thickness were measured either before or after each polymer sample for calibration, and to compensate for daily variations in the pulse energy of the laser system. The x-cut quartz yielded a peak 532 nm energy of 4-6 pJ.
A LabVIEW program was used to automate the Newport ESP 300 motion controller and rotation stage. The sample was rotated from −80 to + 80 degrees in 0.2 degree steps, with 10 pulses averaged per step. Each sample was scanned at least twice, the second scan involvinganother measurement with the sample rotated a quarter turn about the z-axis in an effort to find areas of minimal scattering.
The data were analyzed using a numerical model incorporating the results for birefringent materials reported by Herman and Hayden . Values of the refractive indices n e and no of the polymer samples were measured using a Metricon 2010/M prism coupler. We observed Δn values of 0.001 to 0.003. As a result of the presence of sample wedging and of tiny irregularities in the sample surface, the thickness of the sample at the point of incidence of the beam frequently differed slightly from the measured value of the thickness of the sample as a whole, which was obtained by a Logitech NCG-2 non-contact thickness gauge. For this reason, the value of the sample thickness used in the model was first varied about the measured value over a range of ± 10 μm to optimally align the fringe pattern predicted by theory with that of the data; the value of d 33 was then determined by fitting the theoretical Maker fringe pattern obtained using the “adjusted” sample thickness to the data.
3. Results and discussion
The resulting Maker fringe plots and fits are shown in Figs. 4 through 7 . A representative Maker fringe plot for the x-cut quartz standard is shown in Fig. 8 , indicating a good fit between the data and the theory. This indicates that our Maker fringe experiment is performing properly when a compositionally homogeneous sample with parallel surfaces is examined. In particular, Fig. 4 shows a 1.26 mm thick sample of 10% MNS/PMMA, for which we measured a d 33 of approximately 1.75 pm/V. Figure 5 shows a 1.10 mm thick sample of 15% MNS/PMMA for which we measured a d 33 of approximately 1.50 pm/V. This sample exhibited some signs of aggregation due to the higher chromophore loading, resulting in more scattering of the SHG radiation. Figure 6 shows a 1.22 mm sample of 10% MNA/PMMA, for which we measured a d 33 of approximately 0.7 pm/V. Figure 7 shows a Maker fringe scan of another area of the same 1.22 mm thick sample of 10% MNA/PMMA, for which we measured a d 33 of approximately 0.85 pm/V. While these samples exhibit neither extraordinarily large values of d 33 nor levels of ordering comparable to those observed in DR1 , they do show significant ordering of the chromophores within the polymer host.
Due to a number of sources of error in the measurements, including non-uniformities in chromophore loading, uncertainty in the experimentally measured thickness, poling induced loss , and the possible presence of a wedge angle in the samples, we expect the d 33 values to have uncertainties in range of ± 20%. There are a number of anomalies in the data aside from the presence of aggregates, including the decrease in fringe amplitude in many samples above 60 degrees, and the presence of some SHG signal at zero degrees and at the minima of the fringes. In no case did the quartz standard exhibit this behavior (Fig. 8), so it is reasonable to conclude that it is a property of the poled guest-host samples and not a systematic measurement error in the Maker fringe experiment.
The presence of the 532 nm SHG signal at zero degrees was due to the effects of scattering of both the 1064 nm fundamental and 532 nm harmonic. Our 633 nm scattering measurements performed on each sample before the Maker fringe measurement showed significant scattering from both surface and volume defects. The surface scattering arises from scratches and non-uniformities from electrode impression during poling. The volume effects arise from concentration variations of the chromophore, and they result in concentration banding of the chromophore during the injection molding process. Aggregates consisting of chromophore particulates will also exacerbate scattering, although in cases of low loading (10%) the effects were minimal. Some residual strain-induced birefringence from the injection molding process was also observed in the samples when placed between crossed polarizers. This will also act to disrupt beam propagation, particularly at the point on the sample where the polymer entered the mold and at the edges of the sample where rapid cooling occurred during the molding process.
Thus the 1064 nm fundamental, illuminating a finite portion of the sample surface, encounters these surface and volume irregularities, and a small portion of the 1064 nm fundamental is diffracted or refracted into other angles that allow the SHG phase-matching condition to be satisfied. This causes a small amount of 532 nm harmonic to be generated regardless of the angular orientation of the sample. Therefore this small amount of harmonic gives the appearance of an SHG signal at zero degrees and at the minima of the fringes. Again, this is supported by the lack of any such scattering-generated features in the Maker fringe data for the x-cut quartz standard (Fig. 8).
We feel that the lack of SHG signal above sixty degrees has a similar origin. During the Maker fringe data acquisition, the 1064 and 532 nm beams sweep out many cubic millimeters of illuminated volume, particularly at high angles and on the exit side. This increases dramatically the chances that the beams will encounter concentration banding and the attendant inhomogeneities in refractive index, disrupting the phase matching condition that produces the 532 nm harmonic fringes. Indeed, in our 633 nm scattering measurements we observed that at zero-degree incidence the beam profile is largely intact and exhibits scattering dominated by surface defects. However, at large angles approaching sixty degrees, volume scattering effects are added to the surface scattering and the transmitted 633 nm beam profile begins to break up; the 633 nm beam profile becomes non-distinct at several times the original spot-size. These effects cause the SHG signal to diminish rapidly above sixty degrees.
In addition, the spacing of the fringes is sometimes non-uniform, resulting in some walk-off of the observed fringes from the calculated fringes. We feel this arises from the samples possessing a slight wedge or bow as a by-product of the poling process. In spite of these anomalies, we were able to obtain reasonable fits to most of the data. However, to illustrate the point we show in Fig. 7 calculated Maker fringe patterns for sample thicknesses of 1217 and 1222 μm, a difference of 5 μm, which is half of the ± 10 μm range about the nominal measured value over which we sought the best fit to the fringes. Both have d 33 = 0.85 pm/V. A thickness of 1217 μm provides a good fit for low angles of incidence, but shows misalignment at higher angles; a thickness of 1222 μm (Fig. 7, inset) looks good for angles of incidence in the range 45-60 degrees, but clearly doesn't fit the observed pattern near normal incidence. This clearly demonstrates the presence of non-parallelism in the sample surfaces either in the form of a lenticular distortion or a wedge and that a mere 5 μm variation in thickness over a few square millimeters of sample area can cause misalignment of the observed SHG fringes from the theoretical fit.
To circumvent these problems with surface and volume defects, we are working to improve the concentration uniformity of the chromophore through the use of copolymers and improved polymer processing and molding techniques. We are also working to improve the surface finish and sample surface parallelism through improved electrode and electrode fixture design.
In this paper we described the fabrication and poling of bulk pieces of polymethylmethacrylate (PMMA) 1.2 mm thick doped with various concentrations of 4-methoxy-4ˊ-nitrostilbene (MNS) and 2-methyl-4-nitroaniline (MNA). This accomplishment is a first for samples of this thickness containing chromophores having a transmission window in the visible region of the spectrum. Maker fringe analysis demonstrated that under the poling fields applied, 58 to 67 Vμm−1, substantial ordering of the MNS and MNA chromophores occurred, resulting in birefringence on the order of Δn = 0.001- 0.003 and significant d 33 values. We measured d 33 values in the MNS/PMMA guest-host system of 1.5 to 1.75 pm/V, and 0.7 to 0.85 pm/V in the MNA/PMMA system.
Larger thicknesses of poled electro-optic polymers will result in enhanced performance of electro-optic devices in practical applications, such as those currently being investigated for THz generation. The longer path length available in bulk poled polymers will enable improved THz power, sensitivity, and signal-to-noise ratio in both THz generators and detectors. Large-area Pockels cells constructed from electro-optic guest-host polymers will find new uses in high-speed modulators and shutters, and any application requiring appreciable values of d 33 or r 33 will potentially benefit from bulk-poled guest-host polymers and copolymers. Fabrication and poling of very thick guest-host polymers with greatly improved chromophores and host materials also opens the possibility that bulk-poled guest-host polymer systems may begin to replace bulk inorganic nonlinear crystals in some applications such as second-harmonic generation (SHG) and in optical parametric amplification/generation (OPA/OPG).
We would like to thank Dr. Warren Herman of the Laboratory for Physical Sciences for his most helpful discussions regarding Maker fringe measurement and analysis. We also gratefully acknowledge Kimberly Olver of the U. S. Army Research laboratory for her thickness measurements of the poled polymer samples.
References and links
2. A. Sinyukov and L. M. Hayden, “Efficient electrooptic polymers for THz applications,” J. Phys. Chem. B 108(25), 8515–8522 (2004). [CrossRef]
3. O.-P. Kwon, S.-J. Kwon, M. Jazbinsek, F. D. J. Brunner, J.-I. Seo, C. Hunziker, A. Schneider, H. Yun, Y.-S. Lee, and P. Günter, “Organic phenolic configurationally locked polyene single crystals for electro-optic and terahertz wave applications,” Adv. Funct. Mater. 18(20), 3242–3250 (2008). [CrossRef]
4. C. V. McLaughlin, L. M. Hayden, B. Polishak, S. Huang, J. Luo, T.-D. Kim, and A. K.-Y. Jen, “Wideband 15 THz response using organic electro-optic polymer emitter-sensor pairs at telecommunication wavelengths,” Appl. Phys. Lett. 92(15), 151107 (2008). [CrossRef]
5. L. M. Hayden, A. M. Sinyukov, M. R. Leahy, J. French, P. Lindahl, W. N. Herman, R. J. Twieg, and M. He, “New materials for optical rectification and electrooptic sampling of ultrashort pulses in the terahertz regime,” J. Polym. Sci., B, Polym. Phys. 41(21), 2492–2500 (2003). [CrossRef]
6. A. Natansohn, P. Rochon, C. Barrett, and A. Hay, “Stability of photoinduced orientation of an azo compound into a high-Tg polymer,” Chem. Mater. 7(9), 1612–1615 (1995). [CrossRef]
7. G. Xu, J. Si, X. Liu, Q. G. Yang, P. Ye, Z. Li, and Y. J. Shen, “Comparison of the temperature dependence of optical poling between guest-host and side-chain polymer films,” Appl. Phys. (Berl.) 85, 681–685 (1999).
8. A. Apostoluk, J.-M. Nunzi, V. Boucher, A. Essahlaoui, R. Seveno, H. W. Gundel, C. Monnereau, E. Blart, and F. Odobel, “Permanent light-induced polar orientation via all-optical poling and photothermal cross-linking in a polymer thin film,” Opt. Commun. 260(2), 708–711 (2006). [CrossRef]
9. K. Kitaoka, J. Si, T. Mitsuyu, and K. Hirao, “Optical poling of azo-dye-doped thin films using an ultrashort pulse laser,” Appl. Phys. Lett. 75(2), 157–159 (1999). [CrossRef]
10. K. Kitaoka, N. Matsuoka, J. Si, T. Mitsuyu, and K. Hirao, “Optical poling of phenyl-silica hybrid thin films doped with azo-dye chromophore,” Jpn. J. Appl. Phys. 38(Part 2, No. 9A/B), L1029–L1031 (1999). [CrossRef]
11. R. C. Hoffman, T. M. Pritchett, J. A. Orlicki, J. M. Dougherty, R. H. Lambeth, A. M. Rawlett, W. N. Herman, and D. H. Park, “High-voltage poling of bulk guest-host polymers, chapter 7 in organic thin films for photonics applications,” (ACS Symposium Series, Washington DC 2010).
12. C. Zhao, C.-K. Park, P. N. Prasad, Y. Zhang, S. Ghosal, and R. Burzynski, “Photorefractive polymer with side-chain second-order nonlinear optical and charge-transporting groups,” Chem. Mater. 7(6), 1237–1242 (1995). [CrossRef]
13. V. R. Vangala, B. R. Bhogala, A. Dey, G. R. Desiraju, C. K. Broder, P. S. Smith, R. Mondal, J. A. K. Howard, and C. C. Wilson, “Correspondence between molecular functionality and crystal structures. supramolecular chemistry of a family of homologated aminophenols,” J. Am. Chem. Soc. 125(47), 14495–14509 (2003). [CrossRef] [PubMed]
14. P. D. Maker, R. W. Terhune, M. Nisenoff, and C. M. Savage, “Effects of dispersion and focusing on the production of optical harmonics,” Phys. Rev. Lett. 8(1), 21–22 (1962). [CrossRef]
15. W. N. Herman and L. M. Hayden, “Maker fringes revisited; second harmonic generation from birefringent or absorbing materials,” J. Opt. Soc. Am. B 12(3), 416–427 (1995). [CrossRef]
16. H. Chen, B. Chen, D. Huang, D. Jin, J. D. Luo, A. K.-Y. Jen, and R. Dinu, “Broadband electro-optic polymer modulators with high electro-optic activity and low poling induced optical loss,” Appl. Phys. Lett. 93(4), 043507 (2008). [CrossRef]