This work demonstrates a novel broadband optical switch, based on dynamic-scattering effect in liquid crystals (LCs). Dynamic-scattering-mode technology was developed for display applications over four decades ago, but was displaced in favor of the twisted-nematic LCs. However, with the recent development of more stable LCs, dynamic scattering provides advantages over other technologies for optical switching. We demonstrate broadband polarization-insensitive attenuation of light directly passing thought the cell by 4 to 5 orders of magnitude at 633 nm. The attenuation is accomplished by light scattering to higher angles. Switching times of 150 μs to 10% transmission have been demonstrated. No degradation of devices is found after hundreds of switching cycles. The light-rejection mechanism is due to scattering, induced by disruption of LC director orientation with dopant ion motion with an applied electric field. Angular dependence of scattering is characterized as a function of bias voltage.
© 2016 Optical Society of America
There currently exists a need for optical switches to block intense radiation. The ideal device should have a high on/off contrast, be broadband, polarization-insensitive, sufficiently fast, and durable for the required application Several light switching technologies exist. Some of the common ones based on light absorption are photochromic or electrochromic materials  and twisted nematic liquid crystals  and on light scattering are polymer-dispersed liquid crystal (PDLC) [3,4], holographic polymer–dispersed liquid crystal (H-PDLC) [5,6], polymer stabilized cholesteric liquid crystal (PSCLC) , cycloidal diffractive waveplate in combination with liquid crystal (CDW) [8,9] and dynamic scattering mode (DSM). Most of these technologies are used in devices that detect the light and activate the switching mechanism. All these technologies have advantages and shortcomings and their use depends upon the specific application.
This article reports on a liquid crystal (LC) technology, DSM, for use in light blocking. DSM was initially developed as a display technology in the late 1960s [10–12]. However, in the 1970s a different LC technology, based on twisted nematics , was shown to be superior to DSM for display applications. Research on displays shifted from DSM, and, by the 2000’s, twisted nematic LC technology or an extension of it had been implement for all flat panel displays. The recent development of new, more-stable materials has prompted us to revisit DSM technology for a different application – optical switching. By manipulating light scattering into high angles, DSM optical switches have the advantage of nearly-100% transmission for some wavelengths in the clear state and up to 104-to-105 rejection of on-axis light from visible to near-IR wavelengths in the blocking state. Switching times to the blocking of ~150 μs have been demonstrated.
Schematic diagram of DSM operation is depicted in Fig. 1. DSM cell can be constructed in two ways using a nematic LC. In the first case, the principal orientation of the LC molecules, the director, is perpendicular to the cell windows, Fig. 1(a), referred to as the homeotropic alignment. In the second case, the LC director is parallel to the windows, Fig. 1 (b), forming a homogeneous alignment. When a voltage is applied to the conductive indium-tin-oxide, ITO, film on the windows, ionized dopants in the LC move under the electric field,disrupting the uniformly oriented LC into randomly oriented domains, Fig. 1©. The clear cell now scatters light and appears as frosted glass. When the voltage is removed the cell becomes clear again. Figure 2 shows an example of a DSM cell in its clear Fig. 2(a) and scattering Fig. 2(b) states when illuminated with a 633-nm laser.
Detailed theory of operation of such DSM devices, as developed for display applications, has been described previously [12–19].
The DC dielectric constants, optical densities, and the ion conductivity, parallel and perpendicular to the LC director, are all important parameters for DSM operation. The dielectric constant component normal to the director must be larger than the parallel component, and the ion conductivity component must be higher along the director axis. Under these conditions, two opposing effects on the LC director orientation will occur: the electric field will attempt to force the LC director parallel to the windows while the field-induced movement of ions will force the director towards the perpendicular orientation. The competition between these two forces randomizes the director, leading to scattering of incoming light. For many LCs, the ratio of parallel-to-perpendicular components of ion conductivity is ~1.5 [16,20]. The known parameters for the LCs used in this report are shown in Table 1.
2. Experimental Details
The LC cells consist of conductive indium doped tin oxide (ITO) coated glass, cleaned in H2SO4 and H2O2, rinsed in deionized water, and dried. For homeotropic LC orientation, the ITO surface was coated with polyimide, Merck SE1211, by spinning a solution of the polyimide on the ITO and baking to fix the polyimide in place. For homogeneous LC orientation, the ITO surface was rubbed with lint-free cloth. The cell was assembled in vacuum with degassed LC heated above its nematic-isotropic transition temperature. After assembly the cell was cooled to room temperature over ~1 hr. A 50 μm-thick Mylar sheet was used as a spacer between the cell’s windows. The orientation of the liquid crystal in the homeotropic and homogeneous cells was verified using polarized light.
Two LCs were characterized in this study: n-(4-methoxybenzyliden)-4-n-butylaniline, MBBA, which is a commonly available material from Sigma-Aldrich, and ZLI-4330 LC, supplied by Merck. MBBA, which has been used in the initial DSM cells since the 1960s [23,24], has sufficient impurities so that no dopants were needed to make it conductive. While known to have impurity and stability problems, MBBA is used here as a comparison standard. On the other hand, ZLI-4330 is a purified material, which is normally insulating. Thus, dopants, (2,4,7-trinitro-9-fluoroenylidene)malononitrile, TFM, and n-butylferrocene, BTF, were added to make the material conductive. TFM is an electron acceptor, while BTF is an electron donor. Both compounds can be cycled through their oxidation-reduction states without significant side reactions and with low electrochemical potentials, minimizing electrochemical decomposition of the LC .
Spectroscopic transmission of the cells was performed with a commercial spectrophotometer (Angstrom Sun Technologies, SR500), while optical switching measurements were performed with a 633-nm He-Ne laser. All transmission measurements were referenced to an empty cell. Unless otherwise specified, the transmission was determined by passing laser light through the cell, then onto a 2x2 mm photodiode with a cell-photodiode distance of 45 cm. In this arrangement all on-axis transmitted light is captured, within a solid angle of less than 1 x 10−5 sr. Scattering of light as a function of angle was measured using a large photodiode, 1 cm in diameter, and measuring light intensity as a function of bias voltage for several cell-photodiode separations, as shown below.
Figure 3 shows the transmission of ZLI cells containing just TFM, BTF, and a TFM-BTF mixture. Because of the superior results obtained with the 0.1% combination of TFM and BTF in ZLI, this mixture was used in all experiments described below.
Figure 4 shows the on-axis transmission as a function of bias voltage for LCs ZLI and MBBA. After a few hundred voltage cycles the MBBA degraded and became fixed in the scattering state, only being restored to the clear state after several hours without bias voltage. Others have reported polymer formation on the windows when using MBBA [23,24]. By contrast, we found no degradation with ZLI even after many hundreds of cycles to 150 V.
Figure 5 shows the comparison of the transmission properties of the homeotropic and homogeneous configurations as a function of bias voltage. For static DC bias voltage, both cells show similar transmission for voltages < 60 V, but at higher voltages the homeotropic cell transmission becomes saturated and eventually increases with increasing voltage. The current through the homeotropic cell also saturates, probably due to the buildup of ionized doping gradients in the cell. These gradients can be minimized by using an AC voltage.
The third curve of Fig. 5 was obtained using an AC square-wave voltage at ~3 Hz, which reduces the formation of doping gradients. When applying the AC voltage, the current for the homeotropic configuration increases and the transmission decreases with increasing voltage. Similar transmission curves were obtained with a 60 Hz sine wave voltage. By contrast to the homeotropic cells, homogeneous cells show no current or transmission saturation and minimal change in transmission between DC and AC bias voltages. AC driven cells, both homeotropic and homogeneous, start to attenuate the on-axis light through the cell at ~5 V, while the DC driven cells require higher voltages, ≥ 10 V, for the onset of attenuation.
Figure 6 show the switching times of the homogeneous and homeotropic cells as a function of voltage. The curves are a power-law fit to the data. While the switching times were measured to 660 V the cell cannot withstand this high voltage for more than few tens of milliseconds. Characterization for extended periods of time were limited to 150 V to avoid damage to LC cells.
According to theory, the switching time should be proportional to inverse voltage squared , which is in good agreement with our experimentally-observed trend. Figure 7 shows the transmission as a function of time for a given voltage for the two cells. Similar time-dependence switching is observed for homogeneous cells of ZLI and MBBA and for homeotropic cells of MBBA and ZLI to the extent of our measurements, time and voltages to ~10% transmission.
When a voltage is applied to a homogeneous cell only the ions feel the effect, since the LC’s director is already parallel to the cell’s windows. However, for the homeotropic cell, when a voltage is applied to the LC, the director rotates from perpendicular to parallel orientation relative to the cell’s windows, in addition to the movement of the ions. This combination adds to the disorder of the director. Thus, the homeotropic cell switches faster than the homogeneous cell.
Since the LC is nearly transparent to 633 nm light, light blocking is achieved through scattering. Using a 1-cm diameter photodiode, scattering light intensity as a function of bias voltage was obtained for several separations between the cell and the photodiode, Fig. 8. As shown in Fig. 9, at the smallest separation, 0.1 cm, the light intensity reaching the photodiode for a bias voltage of 150 V decreases only to ~75%. In other words, ~25% of the incoming light is scattered out of a 79° on-axis cone. From the data in Fig. 9 the light intensity per steradian was calculated for several bias voltages and is plotted in Fig. 10. For comparison, a black-body radiator, which has the same radiance when viewed from any angle (Lambert's cosine law), represents the maximum scattering limit and is shown in Fig. 10.
A beneficial property of the DSM is its broadband blocking behavior from 400 to 1700 nm. Figure 11 shows the comparison of an empty cell and a cell filled with ZLI and dopants. Subtracting the empty cell absorption from the measured on-axis light transmission, Fig. 12 was obtained for ZLI and MBBA, showing the transmission of the LC for several bias voltages. Although the cell optical absorption varies with wavelength the light scattering is independent of absorption. The ratio of transmission at 0 V and 50 V as measured by the spectrometer is 10 and 20 at 400 nm and 40 and 120 for ZLI and MBBA respectively. Undoped ZLI is transparent, but after the addition of dopants the liquid becomes tinted light yellow from the BTF. The MBBA is naturally light yellow. To summarize, the transmission spectrum is limited by the combination of the ITO, the glass, the chemistry of the LC, and the dopants.
Below, we discuss possible improvements to the DSM cells operation for light blocking. The switching time of ~150 μs is limited by a ~100 μs dead period where the voltage is applied to the cell, but no reduction in transmission occurs, Fig. 7. Instead of having the director normal to the cell windows, homeotropic, as reported here, a small tilt from normal could eliminate the dead time, potentially reducing the switching time. This can be accomplished by either appropriate chemical treatment of the cell windows or by applying a small AC bias, < 5 V, to the cell. Earlier work has reported homogeneous cell switching faster than homeotropic cells , using LCs available in the 1980s. We speculate that ZLI, having a more negative Δε than MBBA, Table 1, is likely responsible for the superior switching time of the homeotropic cells.
The optical rejection of light can be increased for the homeotropic cell with the addition of a dye to the LC. In the homeotropic cell, a dye that absorbs in the desired spectral region can be aligned with the LC to have its minimal absorption in the clear state. Upon switching to the scattering state the dye would no longer be aligned to minimize its absorption and the overall light rejection by the cell would increase. Similar LC devices, only relying on the dye for light rejection, report clear trans mission of ~90% with blocking transmission of ~10% over a limited optical spectrum [25,26].
The recovery time to the clear state from the scattering state when the scattering bias voltage is stepped to 0 V can be from a few hundred milliseconds to 20 sec. The 50-μm cells discussed here have recovery times depend on the applied voltage from < 1 s for 20 V to 20 s to for 150 V. Thinner cells recover faster with the recovery time decreasing by the square of the cell thickness . Our limited experience found that transmission at a bias voltage (blocking state) is not a strong function of cell thickness, but the bias current increased by more than two orders of magnitude for a cell thickness change from 50 to 10 μm. Additionally, by applying a high frequency voltage on the cell a decrease in clear state recovery time has been reported [10,11,17].
Transmission, as shown in Fig. 12, is a function of the inherent absorption of the LC with its dopants and the dynamic scattering generated by the bias voltage. Figure 13 shows the log base 10 of the difference in transmissions in the clear state, where only the inherent absorption limits transmission, and in the biased state where both inherent absorption and scattering limit transmission. This difference represents the optical density attributed to just the scattering mechanism independent of inherent absorption. Since the optical density slightly decreases with decreasing wavelength, Mie scattering appears to be the dominant mechanism, where the size of the LC scattering domains are larger than the light wavelength.
When the LC cell is biased, turbulence in the LC causes the light to be scattered. The scattering is dynamic, which causes the scattered light to fluctuate in time. This fluctuation is shown in the attached movie where a 2-mm-diameter ~50-mW 780-nm laser beam is passed through a 50-μm-thick homogeneous cell onto a white screen ~60 cm from the cell. The movie shows the transmission as the bias voltage is slowly increased from 0 to 100 V and back to 0 V. Figure 14 is the first frame of the movie, showing the approximate scale and image details.
The best light blocking technology, absorbing or scattering depends upon its detail properties that required for a given application. Details include electrical power requirements, optical bandwidth, switching times, blocking opacity, etc. Light blocking technologies, electrochromics  and LC twisted nematics , can only block light over a comparatively narrow spectrum. Electrochromics only require power to change state while LC twisted nematics require power to remain in either the clear or dark state, and only transmit < 50% of the light in their clear state. The light scattering technologies PDLC [3,4] and H-PDLC [5,6] require power to remain in their clear state. Polymer stabilized scatters PDLC, H-PDLC, and PCLC often have residual light scattering in their clear state. CDW [8,9] and DSM liquid crystal technology do not require power in the clear state. While CDW operates over a limited bandwidth, unlike DSM, but DSM requires higher voltage and longer recovering times to the clear state than most the other technologies. DSM also blocks most of the light that would directly passing through the cell, approaching a Lambersion scattering pattern at high voltages. The above discussion only superficially covers some of the advantages and disadvantages of a few technology in this large field of potential light absorbers and scatters.
The Lincoln Laboratory this work was sponsored by the Defense Advanced Research Projects Agency funded under Air Force Contract FA8721-05-C-0002. The views, opinions, and/or findings expressed are those of the author and should not be interpreted as representing the official views of the U.S. Government. The authors are grateful to Michael Marchant and Sandra Deneault for expert technical assistance.
References and links
1. S. Heusing and M. A. Aegerter, “Sol-gel coatings for electrochromic devices,” in Sol-Gel Processing for Conventional and Alternative Energy, M. Aparicio, A. Jitianu and L. C. Klein, ed. (Springer Science Business Media, 2012).
3. J. L. West, “Phase separation of liquid crystals in polymers,” Mol. Cyst. Liq. Cryst. 157, 427–441 (1988).
4. H. Shim, H.-K. Lyu, B. Allabergenov, Y. Garbovskiy, A. Glushchenko, and B. Choi, “Enhancement of frequency modulation response time for polymer-dispersed liquid crystal,” Liq. Cryst. 5377, 1–7 (2016). [CrossRef]
5. R. L. Sutherland, V. P. Tondiglia, L. V. Natarajan, T. J. Bunning, and W. W. Adams, “Electrically switchable volume gratings in polymer-dispersed liquid crystals,” Appl. Phys. Lett. 64(9), 1074–1076 (1994). [CrossRef]
6. J. Bunning, L. V. Natarajan, V. P. Tondiglia, and R. L. Sutherland, “Holographic Polymer-Dispersed Liquid Crystals (H-PDLCS),” Annu. Rev. Mater. Sci. 30(1), 83–115 (2000). [CrossRef]
7. V. T. Tondiglia, L. V. Natarajan, C. A. Bailey, M. M. Duning, R. L. Sutherland, D. Ke-Yang, A. Voevodin, T. J. White, and T. J. Bunning, “Electrically induced bandwidth broadening in polymer stabilized cholesteric liquid crystals,” J. Appl. Phys. 110(5), 053109 (2011). [CrossRef]
8. N. V. Tabiryan, S. R. Nersisyan, T. J. White, T. J. Bunning, D. M. Steeves, and B. R. Kimball, “Transparent thin film polarizing and optical control systems,” AIP Adv. 1(2), 022153 (2011). [CrossRef]
9. S. R. Nersisyan, N. V. Tabiryan, D. M. Steeves, and B. R. Kimball, “Optical axis gratings in liquid crystals and their use for polarization insensitive optical switching,” J. Nonlinear Opt. Phys. 18(1), 1–47 (2009). [CrossRef]
10. G. H. Heilmeier, L. A. Zanoni, and L. A. Barton, “Dynamic scattering: a new electrooptic effect in certain classes of nematic liquid crystal,” Proc. IEEE 56(7), 1162–1171 (1968). [CrossRef]
11. G. H. Heilmeier, L. A. Zanoni, and L. A. Barton, “Further studies of the dynamic scattering mode in nematic liquid crystals,” IEEE T. Electron Dev. Ed 17(1), 22–26 (1970). [CrossRef]
12. H. Kawamoto, “The history of liquid-crystal displays,” Proc. IEEE 90(4), 460–500 (2002). [CrossRef]
13. H. S. Lim, J. D. Margerum, and A. Graube, “Electrochemical properties of dopants and the D-C dynamic scattering of a nematic liquid crystal,” Elec. Soc. S. 124(9), 1389–1394 (1977). [CrossRef]
14. S. Kai, K. Yamaguchi, and K. Hirakawa, “Observation of flow figures on nematic liquid crystal MBBA,” J. Appl. Phys. 14(11), 1653–1658 (1975). [CrossRef]
15. R. Hérino, “Further investigations and comments on the dynamic scattering mode in nematic liquid crystals,” J. Appl. Phys. 52(5), 3690–3692 (1981). [CrossRef]
16. W. Helfrich, “Conduction-induced alignment of nematic liquid crystals: basic model and stability considerations,” J. Chem. Phys. 51(9), 4092–4105 (1969). [CrossRef]
17. B. Bahadur, “Dynamic scattering mode LCDs,” in Liquid Crystals Applications and Uses, Vol. 1, B. Birendra, ed. (World Scientific Publishing Co. Pte. Ltd., 1995).
18. L. M. Blinov and V. G. Chigrinov, “Modulated and nonuniform structures in nematic liquid crystals,” in Electrooptic Effects in Liquid Crystal Materials, L. M. Blinov and V.G. Chigrinov, ed. (Springer-Verlag New York, Inc., 1994).
19. S. Kai and W. Zimmermann, “Pattern dynamics in the electrodynamics on nematic liquid crystals,” Prog. Theor. Phys. Suppl. 99, 458–491 (1989). [CrossRef]
20. M. Schadt and C. von Planta, “Conductivity relaxation in positive dielectric liquid crystals,” J. Chem. Phys. 63(10), 4379–4383 (1975). [CrossRef]
21. “Physical properties of liquid crystals,” in Springer Handbook of Condensed Matter and Materials Data, Werner Martienssen, and Hans Warlimont, ed. (Springer Science & Business Media, 2006).
22. B.-X. Li, V. Borshch, S. V. Shiyanovskii, S.-B. Liu, and O. D. Lavrentovich, “Electro-optic switching of dielectrically negative nematic through nanosecond electric modification of order parameter,” Appl. Phys. Lett. 104(20), 201105 (2014). [CrossRef]
23. S. Barret, F. Gaspard, R. Herino, and F. Mondon, “Dynamic scattering in nematic liquid crystals under dc conditions. I. Monitoring of electrode processes and lifetime investigation,” J. Appl. Phys. 47(6), 2375–2377 (1976). [CrossRef]
24. S. Barret, F. Gaspard, R. Herino, and F. Mondon, “Dynamic scattering in nematic liquid crystals under dc conditions. II. Monitoring of electrode processes and lifetime investigation,” J. Appl. Phys. 47(6), 2378–2381 (1976). [CrossRef]
25. H. V. Ivashchenko and V. G. Rumyantsev, “Molecular Crystals and Liquid Crystals Dyes” in Liquid Crystals, G. H. Dienes, ed. (Gordon and Breach Science Publishers S.A., 1987).
27. L. T. Creagh, A. R. Kmetz, and R. A. Reynolds, “Performance characteristics of nematic liquid crystal display devices,” IEEE Trans. Electron Dev. ED-18(9), 672–679 (1971). [CrossRef]