High-speed photorefractive polymer composites D. Wright, M. A. Dı ´ az-Garcı ´ a, J. D. Casperson, M. DeClue, and W. E. Moerner a) Department of Chemistry, University of California San Diego, La Jolla, California 92093-0340 R. J. Twieg Department of Chemistry, Kent State University, Kent, Ohio 44242 Received 22 June 1998; accepted for publication 10 July 1998 Two photorefractive polymer composites are presented that exhibit the fastest response times reported to date by an order of magnitude (  g 5 ms at 1 W/cm 2 ), while maintaining large gain coefficients ( 230 and 130 cm -1 ). These materials show promise for video-rate optical processing applications. The factors limiting the photorefractive speed in these materials are investigated. © 1998 American Institute of Physics. S0003-69519803437-8 In the years since photorefractivity was demonstrated in polymers for the first time in 1991, 1 there have been many advances in material performance. The first materials 2 dis- played gain coefficients (  ) 1 cm -1 , no net gain, diffrac- tion efficiencies (  ) 10 -6 , and response times (  g ) of many minutes at 1 W/cm 2 . In 1993 a milestone was achieved with a composite of PVK/FDEANST/TNF. 3 At 753 nm this material displayed a net 7 cm -1 ,  10 -3 , and  g 1 s. The following year a composite of PVK/DMNPAA/ECZ/ TNF passed another milestone by exhibiting a net  200 cm -1 , overmodulated  in 100  m thick samples, and  g 100 ms. 4 Today, several materials have exhibited performance similar to the DMNPAA composite. 5 In terms of speed, the fastest material reported has been a composite based on polysilane which at 647 nm had a  g 40 ms, although no net gain. 6 In this letter we present polymer com- posites with  g as low as 4 ms, well into the video response range, that also have  100 cm -1 . The polymer composites consisted of the hole transport- ing polymer polyn-vinyl carbazolePVK, doped with a nonlinear optical chromophore NLO35 wt %, the liquid plasticizer butyl benzyl phthalate BBP15 wt %, and C 60 0.5 wt %, which is used as a sensitizer. The chemical struc- tures of the different NLOs used are given in Fig. 1. Com- posites were sandwiched between two indium–tin–oxide ITO coated glass plates to yield samples with thickness between 60 and 100 m as described previously. 7 The photorefractive properties of these samples were de- termined using a standard two wave mixing TWM setup. 7 The tilted geometry configuration was employed, in which two equal intensity p-polarized 647 nm beams intersected in the material at external angles of 30° and 60° with respect to the sample normal. The phase of the pump beam with respect to the signal was controlled by moving a mirror with a pi- ezoelectric actuator. A typical run consisted of the following steps. First, an electric field was applied to the sample with the beams off. After some time, the pump beam was turned on for 1 s, and the transmitted power was monitored with time to check for the presence of beam fanning. 8 Next, the photoconductivity (  ph ) was measured by the following method. Both beams were turned on for 200–500 ms while driving the piezo mir- ror with a 1 kHz triangle wave. This caused the interference pattern of the two beams to translate back and forth over many grating periods faster than a grating could be written, so the sample would respond as if it were illuminated with incoherent light. The current through the sample with respect to time was measured before and during this incoherent illu- mination to obtain the dark and photocurrent, which were converted to dark photoconductivity (  d ) and  ph . Next, without driving the piezo mirror, both beams were turned on and the transmitted powers were measured with respect to time. An example of this measurement for the gain beam is shown in Fig. 2 the fits in this figure that are used to obtain  g are explained below. After steady state energy transfer was achieved, the gain  was measured to determine . Finally, the output intensities were monitored while the piezo mirror was used to quickly translate the light pattern with respect to the index grating. The resulting intensity oscilla- tions were used to obtain the grating phase shift and the maximum index modulation (  n ). In the simplest single carrier model of photorefractivity the gain transients are exponential. 9 However, in polymers the transients are usually nonexponential and must be fit with some other function. In the past we and other groups have used the sum of two exponentials and reported only the early time  g obtained from the fit. In this work we have used this technique to obtain  g that can be compared to other mate- rials in the literature. In addition, we have also fit our tran- sients with a stretched exponential function: a Electronic mail: w.e.moerner@standford.edu FIG. 1. NLO chemical structures. APPLIED PHYSICS LETTERS VOLUME 73, NUMBER 11 14 SEPTEMBER 1998 1490 0003-6951/98/73(11)/1490/3/$15.00 © 1998 American Institute of Physics Downloaded 03 Jul 2002 to 193.145.234.37. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp