March 15, 1996 / Vol. 21, No. 6 / OPTICS LETTERS 429

Red-band holographic storage in azo dyesensitized by noncoherent light

Pengfei Wu, Wenju Chen, Xiong Gong, Guilan Zhang, and Guoqing Tang

Institute of Modern Optics, Nankai University, Tianjin 300071, China

Received September 29, 1995

Biphoton holographic storage recorded with 633-nm He –Ne beams has been investigated in Methyl Yellow-doped polystyrene film sensitized by weak noncoherent light from a low-pressure Hg lamp. It is demonstratedthat the holography in the sample is erasable without fading. The dynamic behavior of the hologram isdependent on the cis– trans isomerization by noncoherent light and 633-nm beams and photoinduced heat. 1996 Optical Society of America

Organic photorefractive and photochromic materials,used in holography and degenerate four-wave mix-ing, have attracted attention in the past few yearsbecause of their potential applications in optical in-formation storage and processing.1,2 Organic mate-rials have some advantages over inorganic crystals,e.g., very high exposure sensitivities, fast dynamics,ease of preparation, and low cost. Unfortunately theirabsorption bands lie mostly in the blue-green wave-length or shorter regions, which are not compatiblewith miniature diode lasers.3 This limitation pro-hibits these materials from being used in integratedoptics. Recently biphoton holography was reportedand realized in the extension of the recording wave-length in organic materials,4 – 7 but this method is notcurrently suitable for practical use because of a too-short recording wavelength,4,5 photoinstability,6 andlack of erasability.7 In this Letter we report the studyof erasable biphoton holography, which is recordedwith He–Ne lasers in azo dye sensitized by a weak non-coherent light, and we explain the photophysical andphotochemical dynamic processes of optical storage anderasure.

The sample used in our experiments is MethylYellow (MY) as 4-dimethylaminazobenzene, which isdoped in a polystyrene (PS) polymer matrix. Thepreparation of MY-PS film is as follows: MY and PSare dissolved separately in chloroform, and then theMY solution is poured into the PS solution. Afterbeing stirred for 1 h, the mixed solution is coatedonto a cleaned glass slide and allowed to dry at roomtemperature in a clean, dry cabinet. The contentof MY in the PS matrix is 1% by weight, and thethickness of the film is approximately 10 mm. Theexperimental setup for biphoton holographic memory isschematically shown in Fig. 1. The 633-nm coherentradiation from a He–Ne laser is split into threebeams by beam splitters (BS1, BS2) and mirrors (M1–M3). Beams 1 and 2 intersect at an angle of 5± andare used to write a hologram in the sample; beam 3counterpropagates with beam 1 and reads the grating.The three beams have equal powers and at the samplehave spot areas of 5 mm2. The same region on thesample is simultaneously irradiated by a focused short-wave noncoherent beam from a low-pressure Hg lampwith an intensity of 8 mWycm2. A diffractive signal

0146-9592/96/060429-03$6.00/0

(beam 4) is split off by beam splitter BS3, detected by aphotomultiplier tube (PMT), and sent into an amplifier(AMP) and a data-acquisition system (DAS). Thisarrangement is also fitted to degenerate four-wavemixing, so the real-time biphoton holography can beregarded as biphoton degenerate four-wave mixing.

We find that no diffractive signal is detectedby the PMT with only the irradiation of the three633-nm beams, but the signal will emerge when thelow-pressure Hg lamp is also on. The first experimentis to determine the temporal behavior of the diffractionefficiency as a function of the exposure configurationof the noncoherent light and the intensity of 633-nmlight. The results are shown in Fig. 2. At the low-power levels of the 633-nm beam, the diffractive signalis very weak in the beginning (t ­ 0–2 min) while thesample is simultaneously exposed to the noncoherentlight and the three 633-nm beams. However, at timet ­ 2 min, when the noncoherent light is blocked, thediffraction efficiency begins to grow with time andfinally reaches a saturable value. As the intensity of633-nm light increases, the diffraction efficiency inthe time t ­ 0–2 min is also enhanced; and at timet . 2 min, when the maximum diffraction eff iciency isachieved, the signal begins to decay slowly. The riseand the decay of the diffractive signal become fast withthe increase of the 633-nm intensity. The diffraction

Fig. 1. Experimental setup of biphoton holographic stor-age and erasure.

1996 Optical Society of America

430 OPTICS LETTERS / Vol. 21, No. 6 / March 15, 1996

Fig. 2. Temporal development of the diffractive eff icien-cies at different 633-nm intensities. At t ­ 0 min, both633-nm and Hg light are on; at t ­ 2 min, Hg light is off.

efficiency achieves the maximum in the case of 633-nmintensity of ,20 mWycm2 in each beam.

The dynamics of biphoton holographic memory de-pends on the cis,trans isomerization of azo moleculesby light and heat. At room temperature the stableconfiguration of the azo molecule is the ground state(S0) of the trans-isomer whose absorption peak in thevisible region is near 400 nm and transparent to wave-lengths longer than 520 nm.8 However, when thetrans-isomer is exposed to short-wavelength light (hv1;e.g., Hg light), it may be excited to its excited states (S1and T1), and then it changes to a cis-isomer. The ab-sorption of the cis-isomer is stronger than that of thetrans-isomer in the long-wavelength region (approxi-mately 500 nm) and up to 633 nm. The cis-isomer canreturn back to the trans-isomer from its excited states(S10 , T10 ) by absorbing the long-wavelength light (hv2;e.g., He–Ne laser), and it also relaxes thermally to thetrans-isomer from ground states (S00 ! S0) with thelifetime (t) of the cis-isomer. The photophysical andphotochemical processes are illustrated in Fig. 3.

Since the photoisomerization between cis and transisomers occurs in their excited states and depends onthe population of the excited states, the overall rate (ki)of photoisomerization is proportional to the light inten-sity sIid, the absorption cross section ssid, and the quan-tum yield sfid (Ref. 4): ki ­ sifiIiyhvi, where hvi isthe photon energy and i ­ 1, 2. For the fixed inten-sity of noncoherent light, the trans ! cis photoisom-erization rate si ­ 1d is constant, but the cis ! transphotoisomerization rate si ­ 2d is varied as the in-tensity modulation of the 633-nm writing interferencefield. Hence the spatial distribution of two isomers inthe irradiated area of the sample is also periodicallymodulated, which forms an absorption and refractive-index grating for the 633-nm read beam. When the633-nm beam power is lower than that of Hg light, thehomogenous noncoherent radiation will bury the inten-sity modulation of writing beams (erasable effect) sothat the spatial difference of photoisomerization willbe obscure and the diffraction efficiency will be veryweak. When the noncoherent light is blocked, the cis-isomers located in bright fringe will continue to absorb633-nm light and isomerize to the trans-isomers be-cause of their long lifetime, so the spatial populationmodulation between cis and trans is enhanced. Al-though the accumulation of this isomerization grating

increases slowly because of the small rate of cis ! transisomerization at weak 633-nm intensity, the diffrac-tion efficiency still reaches a stronger steady value.The reason is that the weak photoinduced heat effectleads the molecules to a lack of activity and the cis-isomers in the dark fringe are effectively bound by therigid PS polymer. However, with the increase of the633-nm intensity, the erasable effect of homogeneousnoncoherent light is weak compared with the strong in-terference modulation of 633-nm light. Consequentlythe diffraction efficiency in the beginning is enhanced.When the noncoherent light is blocked, the photo-isomerization of the trans ! cis direction disappearsand the cis-isomers in the bright fringe isomerize fur-ther to the trans form at a faster rate sk2 ~ I2d; i.e., theresumption of isomerization grating becomes fast. Onthe other hand, an increase in the photoinduced heateffect weakens the binding of the PS polymers to cis-isomers and hastens their thermal relaxation. Thestrong read beam takes the cis-isomers in dark fringeback to the transform as well. As a result, the diffrac-tion eff iciency will decay more rapidly after the maxi-mum is reached.

The second experiment is to investigate the era-sure of gratings. In the case of strong 633-nm light(200 mWycm2 in each beam), we present three kindsof erasure procedures (at t ­ 2 min), as shown inFig. 4: (a) Blocking the Hg lamp only. As describedabove, the grating is automatically erased fromthe peak by both photoinduced heat and intense633-nm light. The rate of erasure is determined by

Fig. 3. Energy-level and isomerization procedures of theazo molecule with two isomers.

Fig. 4. Grating erasure with a 633-nm intensity of200 mWycm2 in each beam.

March 15, 1996 / Vol. 21, No. 6 / OPTICS LETTERS 431

Fig. 5. Long-time grating formation and erasure with a633-nm intensity of 5 mWycm2 in each beam.

the cis ! trans photoisomerization rate (relativelyfast time scale) and thermal relaxation (relativelyslow time scale.) (b) Simultaneously blocking writingbeams and noncoherent light. Figure 4 shows thatthe signal does not grow but declines directly becauseno gratings build up in the absence of write beams;additionally, the single strong 633-nm read beammakes the grating vanish more quickly than in case(a). (c) Shutting off the write beams only. This caseis similar to case (b), but the erasure of the gratingis much quicker because the gratings are erased byboth noncoherent light and the read beam. In thecase of the weak 633-nm intensity, the situation isquite different. Figure 5 shows the procedure of longgrating formation and erasure with a 633-nm intensityof 5 mWycm2 in each beam: At time t ­ 0–1 min, thesample is simultaneously exposed to the 633-nm lightand the noncoherent light. Then, at t ­ 1 min, the Hglight is blocked. After the grating has reached nearsaturation (t ­ 11 min), the write beams are turnedoff. The experimental result shows that the signaldecays very slowly (long-time storage, which can keepfor more than one day). If the Hg light is turnedon again at t ­ 15 min, the grating will be erasedrapidly to zero because of the renewed isomerizationby homogenous noncoherent light. At t ­ 18 min thisprocedure is repeated again in the same region of thesample without any bleaching.

In conclusion, we have performed reversible biphotonholographic storage in a polymer layer doped with azodye. The dynamic behavior of diffraction efficiency isgoverned by the exposure configuration of noncoherentlight and the intensities of the write and read beams.We have shown that the weak 633-nm beams aresuitable to long-time storage; otherwise the strong633-nm light brings about a relatively rapid timeresponse. Moreover this biphoton storage needsonly a weak noncoherent preexcitation (even as lowas 1 mWycm2) in MY-PS film in order to record ahologram with 633-nm laser beams. Eventually therecording wavelength can be extended because thediffraction efficiency and degenerate four-wave mixingref lectivity have higher values at wavelengths in thewings of the absorption spectrum.9 In the next step,we hope to apply the infrared diode laser to biphotonstorage by using new azo dyes with faster responsetime, higher eff iciency, and longer storage time.

This research was supported by a grant from the KeyProject of State Major Basic Research of China.

References

1. W. Nie, Adv. Mater. 5, 520 (1993).2. K. K. Sharma, K. D. Rao, and G. R. Kumar, Opt.

Quantum Electron. 26, 1 (1994).3. T. Tamir, Integrated Optics (Springer-Verlag, Berlin,

1979), p. 271.4. V. Weiss, A. A. Friesem, and V. A. Kronganz, Opt. Lett.

18, 1089 (1993).5. G. C. Bjorklund, C. Brauchle, D. M. Burland, and D. C.

Alvarez, Opt. Lett. 6, 159 (1981).6. R. M. Tarkka, M. E. Talbot, D. J. Brady, and G. B.

Schuster, Opt. Commun. 109, 54 (1981).7. D. J. Lougnot, D. Ritzenthaler, C. Carre, and J. P.

Fouassier, J. Appl. Phys. 63, 4841 (1988).8. J. F. Rabek, Photochemistry and Photophysics (CRC,

Boca Raton, Fla., 1990), pp. 120–141.9. C. Egami, K. Nakagawa, and H. Fujiwara, Jpn. J. Appl.

Phys. 31, 2937 (1992).