5146

OPTICS LETTERS / Vol. 39, No. 17 / September 1, 2014

Negative dispersion of birefringence in two-dimensionally self-organized smectic liquid crystal and monomer thin film Hyojin Lee and Ji-Hoon Lee* Division of Electronics Engineering, Chonbuk National University, Jeonju, Jeonbuk 561-756, South Korea *Corresponding author: [email protected] Received July 1, 2014; revised July 22, 2014; accepted July 23, 2014; posted July 29, 2014 (Doc. ID 214952); published August 26, 2014 We suggest a method to obtain a negative dispersion (ND) of birefringence using a two-dimensional self-organization of smectic liquid crystal (LC) and monomer molecules. The averaged orientation of the smectic LC was the layer normal direction with the extraordinary refractive index ne . Meanwhile, the orientation of the monomer molecules was templated by the host-smectic LC and parallel to the layer plane corresponding to the ordinary refractive index no . We selected the LC molecules absorbing a shorter wavelength of UV light rather than the polymerized monomers, hence ne was more smoothly decreased than no in the visible-wavelength range. Consequently, the birefringence Δn ≡ ne − no was increased with a longer wavelength, thus giving a ND of birefringence. Using the proposed method, the ND of birefringence could be obtained in a single layer, which is desirable for thin flexible applications. © 2014 Optical Society of America OCIS codes: (160.1190) Anisotropic optical materials; (160.3710) Liquid crystals; (310.6860) Thin films, optical properties. http://dx.doi.org/10.1364/OL.39.005146

0146-9592/14/175146-04$15.00/0

wavelength of light λa of the material. As λ is further away from λa , the magnitude jdn∕dλj is decreased, showing less dependence on λ. Therefore, λa of ne should be shorter than that of no to obtain the ND property. In the copolymer single-layer approach [10–12], the ND property of the birefringence was achieved by copolymerization of different chemical moieties which play a role of ne and no . ne was rapidly decreased in the visible wavelength range, while no was slowly decreased, as shown in Fig. 1(b). In the reactive mesogen approach [13,14], ND was obtained by synthesis of a ladder-shaped reactive mesogen whose longitudinal units play a role of ne and connecting rods play a role of no . λa of the longitudinal unit was shorter than that of the connecting rods, and consequently, the ND dispersion property was shown. Although both methods showed the ND property, the copolymer approach is difficult to thin and a stretching

(b)

(a) ne

ne n

n

Generally, the birefringence Δn of natural materials decreases with a longer wavelength of light λ, which is called a positive dispersion (PD) of Δn [1]. Δn is defined as ne − no , where ne and no are the extraordinary and ordinary refractive index of the medium, respectively. The retarder made of the PD materials inevitably shows a change of the phase retardation Γ, depending on λ. Here, Γ is given by 2πΔnd∕λ, where d is the thickness of the retarder. Thus, the typical bandwidth of the PD retarder is generally very narrow, about 5–10 nm. In particular, the retarder films with PD used in display devices for the viewing angle compensation of liquid crystal (LC) display or the antireflection film of an organic lightemitting diode result in degradation of image quality due to the mismatch of Γ for different λ [2,3]. To avoid the λ dependence of Γ, a retarder showing an increase of Δn with a longer λ, i.e., negative dispersion (ND) of Δn, has been developed [4–12]. The most wellknown ND retarder is the zeroth-order waveplate composed of multilayer PD retarders whose optic axes were orthogonally or obliquely aligned [4–7]. The dispersion of Γ after passing through the first retarder is compensated by the following retarders. Recently, ND using adiabatic rotation of the polarization of light was also reported using multilayers of twist-oriented reactive mesogen [8,9]. However, it is more desirable that the ND retarder be fabricated using a single layer for flexible application and lower cost. Several reports of the single-layer ND retarder have recently been published using copolymers [10–12] or reactive mesogen [13,14]. The physics of ND in the single layer is based on the different dispersion properties of ne and no as illustrated in Fig. 1. The PD materials show a rapid decrease of ne and a slow decrease of no , resulting in a decrease of Δn with longer λ [Fig. 1(a)]. On the other hand, ne and no of the ND material has a reverse λ dependence, i.e., a slow decrease of ne and a rapid decrease of no , thus giving an increase of Δn with longer λ [Fig. 1(b)]. The slope dn∕dλ is related to the absorption

no

n decreasing with longer

no

n increasing with longer

(c) no (Long UV absorption)

ne (Short UV absorption)

Smectic LC Layer

Monomers

Fig. 1. Schematic illustration of the wavelength λ dependence of ne and no in the (a) positive and (b) negative dispersion of Δn medium. (c) Suggested 2D self-organization of the smectic LC and monomers showing negative dispersion of Δn. © 2014 Optical Society of America

September 1, 2014 / Vol. 39, No. 17 / OPTICS LETTERS

process is required for the film application. The laddershaped mesogen approach is easy to thin about several micrometers and can be fabricated with a simple coating process. However, the ladder-shaped mesogen is difficult to synthesize and the material showed LC phase in a very narrow range (about 3 K) due to its crystallinity, making the fabrication process difficult. In this Letter, we suggest an approach to achieving the ND dispersion using a 2-dimensional (2D) selforganization of smectic LC and monomers [15–19]. We used the smectic LC as a template to induce the anisotropic orientation of the monomers, as reported previously [17–19]. We note that the LC showed PD before the self-organization, and the monomer has no birefringence. The host smectic LC formed a layer structure, where the LC molecules were aligned to the layer normal direction [Fig. 1(c)]. The refractive index to the layer normal direction corresponds to ne whose dispersion property is determined by LC. The reactive monomers are not liquid crystalline [Fig. 2(a)] and are thus separated from LC and located at the interlayer space parallel to the layer plane. The refractive index to the layer-plane direction corresponds to no whose dispersion is predominantly affected by the monomers. We selected LC whose λa is further away from the visible wavelength of light [Fig. 2(b)] than that of the polymerized monomers. Consequently, ne was smoothly decreased in the visible wavelength range, while no was rapidly decreased in the same range. Consequently, the birefringence Δn
ne − no was increased with a longer wavelength, thus showing ND of Δn. The retarder film suggested in this

Absorption (arb. units)

(b) 1.0 0.8 0.6

Pure LC

0.4

UV-polymerized monomers 0.2 0.0 300

400

500

600

700

800

(nm) Fig. 2. (a) Chemical structure of the monomers and photoinitiator. (b) Absorption spectra of the pure LC and the UV-polymerized monomers–photoinitiator mixture.

5147

Letter requires neither a complex synthesis process nor a stretching process, which is more desirable for the mass production of the compensation film. A commercial polyimide PIA-X189-KU1 (JNC) was coated onto a glass substrate and then baked at 200°C for 1 h. The substrates were rubbed with a cotton cloth and assembled in an antiparallel fashion. The cell gap d was maintained to be 2.3  0.2 μm using bead spacers. Commercial smectic LC mixture FELIX015-100 (Clariant) was used as a host and mixed with a mixture of reactive monomers triallyl-1,3,5-triazine-2,4,6(1h,3h,5h)trione (triallyl, Aldrich), 1,6-hexanediol diacrylate (HDDA, Aldrich), and photoinitiator Irgacure651 (Ciba Chem) [Fig. 2(a)]. The phase sequence of pure LC is Cr -9°C SmC 67°C SmA 82°C N 87°C I. The weight fraction of the photoinitiator was kept at 5 wt. % to the total amount of the monomers through all experiments. The LCmonomer mixture was injected into the empty sample by a capillary action at 100°C. Then the sample was cooled to 25°C and a UV light of 30 mW∕cm2 was irradiated for 1 m. To check λa of the materials, the LC and monomers were filled in the polymethylmetacrylate (PMMA) cuvette and the absorption intensity was measured using a SV2100R (K-MAC) UV-vis. spectrometer. The retardation α ≡ Δnd was measured using a commercial retardation measurement instrument, Axo Scan-OPMF2 (Axo Metrics), utilizing the Muller matrix method. The layer spacing l was checked from the small-angle x-ray scattering (SAXS) spectrum using a D8-Discover (Bruker) [20]. λ of the x-ray source was 0.154 nm and the temperature was maintained at 25°C during the measurement. For the SAXS measurement, the LC-monomer mixture was injected into a quartz capillary tube with a diameter of 0.5 mm. To identify the orientation of the monomers, a IR dichroism technique was used [20]. The IR absorption intensity was measured versus the polarization direction of the IR light using a Fourier transform-IR (FT-IR) spectrometer (FTS7000, Biorad). For the IR absorption measurement, CaF2 slides were used as substrates whose surface was treated with the same conditions above. Figure 3(a) represents the retardation α of the pure LC and LC-monomer mixture samples in the visible-light region at 25°C. α of the pure LC was decreased with longer λ, showing PD. With a greater concentration of the monomers, the magnitude of α was decreased due to the increased fraction of the optically isotropic monomers. Interestingly, the PD property was converted to ND over 34 wt. % of the monomers. Figure 3(b) shows αλ normalized to the α value at 550 nm, α550. It is clearly seen that the dispersion property was gradually inverted from PD to ND with a greater concentration of the monomers. The black line denoted with an arrow in Fig. 3(b) corresponds to the ideal-dispersion curve, which gives the same Γ through all wavelengths of light. For this ideal dispersion, α450∕α550 and α650∕α550 should be 0.82 and 1.18, respectively. α450∕α550 and α650∕α550 of the 38 wt. % monomer sample were 0.38 and 1.14, respectively. The dispersion was nearly similar to the ideal one for λ > 480 nm, but was deviated for λ < 480 nm. To confirm the suggested model for the observed ND property, we investigated the constituent molecules orientation using SAXS [Fig. 4(a)] and the IR-dichroism

(a)

OPTICS LETTERS / Vol. 39, No. 17 / September 1, 2014 Monomer 0 wt% 13 wt% 22 wt% 34 wt% 38 wt% 42 wt%

400 350

(nm)

300 250 200 150 100 50 0 450

500

550

600

(a) Intensity (arb. units)

5148

1.0 0.8 0.6 LC+Monomers 0.4

Pure LC

0.2 0.0

650

2

4

(nm) 2.0

Monomer 0 wt% 13 wt% 22 wt% 34 wt% Ideal retarder 38 wt% 42 wt%

1.8

( / 550)

1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 450

500

550

8

10

2 deg.)

600

650

(nm) Fig. 3. (a) Retardation α versus λ for the pure LC and LC– monomer mixture samples with various concentrations of the monomers at 25°C. (b) αλ normalized to α value at λ
550 nm, α550.

technique [Fig. 4(b)]. The pure LC showed a SAXS resonance peak at 2θ
3.51°, which corresponds to the layer spacing l
2.52 nm [Fig. 4(a)]. On the other hand, the LC mixed with 38 wt. % monomers showed the resonance peak at smaller angle 2θ
3.39°, which corresponds to l
2.60 nm. The increased l implies the monomers were located at the interlayer space [Fig. 1(c)]. Next, we examined the IR dichroism data to find the orientation of the monomers [Fig. 4(b)]. The absorption of the biphenyl stretching of the LC molecule at 1600 cm−1 was the strongest when the polarization direction of the IR light was parallel to the layer normal direction. On the other hand, the strongest absorption of the C═O bond of the triallyl molecule at 1690 cm−1 was shown when the polarization direction of the IR light was parallel to the layer plane. This result indicates the polymerized monomers were aligned parallel to the layer plane [Fig. 1(c)]. Figure 5 shows the polarizing optical microscopy (POM) image of the samples with various monomer concentrations. When the rubbing direction was at 0° to the polarizer, the pure LC showed some light leakage due to the azimuthal tilt of the LC molecules [Fig. 5(a)]. However, with a greater fraction of the monomers, the sample showed a darker state when the rubbing direction was parallel to the polarizer. As described in the literature [21], the smectic LC molecules were tilted with an anticlinic fashion when the interlayer space was segregated with the monomer molecules. Consequently, the optic

(b) 1.0 Absorption (arb. units)

(b)

6

0.9

120

Layer parallel 90 60

-1

LC 1600 cm -1 Monomer 1520 cm

30

150

0.8 180

0 Layer normal // Rubbing

0.8 0.9

330

210

1.0 240

300 270

Fig. 4. (a) SAXS intensity of the pure LC and 38 wt. % monomers-mixed LC versus 2θ. (b) IR absorption intensity of the biphenyl stretching of LC at 1600 cm−1 and C═O bond of triallyl at 1690 cm−1 versus rotation angle of the polarization direction of the input IR light. The data were taken at 25°C.

axis became more parallel with the rubbing direction. With the monomers concentration over 34 wt. %, the rubbing direction of the sample was nearly the same as the ne direction with a variance less than 2°. We should note that the monomer domains were parallel to the rubbing direction when the monomer concentrations were less than 27 wt. % (dark area in [Figs. 5(h)–5(j)]). However, with a greater concentration of monomers over 27 wt. %, the fraction of the monomer domains parallel to the layer plane increased more [Figs. 5(k)–5(n)] and clearly predominant in the sample with 42 wt. % monomers [Fig. 5(n)]. As described above, we selected the monomers whose λa was longer than that of LC [Fig. 2(b)]. From the experimental results of Figs. 4 and 5, it was confirmed that more monomers with longer λa were located at the interlayer space and oriented parallel to the layer plane with a greater fraction of the monomers. The layer-plane direction corresponds to the ordinary axis with no . Meanwhile, the LC molecules with shorter λa were nearly aligned to the layer normal direction when the monomer concentration was over 34 wt. %. Because the monomers that were aligned to the no direction absorbed a shorter wavelength of light than the LC molecules that were aligned to the ne direction [Fig. 2(b)], no might decrease more rapidly in the visible wavelength range compared to ne [Fig. 1(b)],

September 1, 2014 / Vol. 39, No. 17 / OPTICS LETTERS

5149

Fig. 5. POM images of the samples with various monomer concentrations (a), (h) 0, (b), (i) 13, (c), (j), 22, (d), (k), 27, (e), (l), 34, (f), (m), 38, and (g), (n) 42 wt. %. (a)–(g) were taken when the rubbing direction was parallel to the polarizer. (h)–(n) were taken when the rubbing direction was at 45° to the polarizer. The pictures were taken at 25°C. Scale bars correspond to 100 μm.

resulting in the ND property. The promotion of ND with a greater concentration of the monomers in Fig. 3 could be also understood provided more fraction of the monomers were aligned to the no direction, enhancing rapid decrease of no . We investigated the temperature dependence of the ND property of the 38 wt. % monomer sample. The ND property gradually decreased with increasing temperature and reverted to PD over 57°C. The dα∕dλ continuously decreased up to nematic-isotropic (NI) phase transition temperature. The decreased ND property is presumably related to the increased longitudinal molecular fluctuation deforming the layer-parallel orientation of the polymer networks [21]. The detailed temperature dependence of the ND property and the change of the molecular orientation are under investigation and will be announced in the near future. We also checked the contrast ratio and response time of the 38 wt. % monomer sample. The contrast ratio was 130∶1 and saturated at 10 V. Both the rising and falling time was less than 0.5 ms through the whole gray-level switching. Thus, the suggested system showed good electro-optical performance even in the dynamic mode. To summarize, we suggested a method to obtain ND of birefringence using a 2D self-organization of the smectic LC and monomers. The polymerized monomers with longer λa were aligned to the no direction, and no might be rapidly decreased with longer λ in the visible wavelength range. On the other hand, the LC molecules with shorter λa were aligned to the ne direction, thus resulting in a slow decrease of ne . Consequently, Δn was increased with longer λ, showing ND of birefringence. The method suggested in this Letter used conventional PD LC and nonbirefringence monomers. Thus, the proposed method does not require a complex synthesis process and is expected to be useful for the mass production of ND retarder film. This research was supported by the Basic Science Research Program through the National Research

Foundation of Korea (NRF) and funded by the Ministry of Science, Information and Communication Technology (ICT) and Future Planning (NRF2013R1A1A1058681). This work was also supported by the Brain Korea 21 PLUS Project. References 1. H. Mattoussi, M. Srinviasarao, P. Kaatz, and G. C. Berry, Mol. Cryst. Liq. Cryst. 223, 69 (1992). 2. P. Yeh and C. Gu, Optics of Liquid Crystal Displays (Wiley, 1999), pp. 357–384. 3. B. Bahadur, Mol. Cryst. Liq. Cryst. 109, 3 (1984). 4. S. Pancharatnam, Proc. Ind. Acad. Sci. A 41, 130 (1955). 5. D. Clarke, Opt. Acta 14, 343 (1967). 6. P. Hariharan, Opt. Eng. 35, 3335 (1996). 7. Y.-C. Yang and D.-K. Yang, Proc. SID Digest 39, 1955 (2008). 8. S. Shen, J. She, and T. Tao, J. Opt. Soc. Am. 22, 961 (2005). 9. R. K. Komanduri, K. F. Lawler, and M. J. Escuti, Opt. Express 21, 404 (2013). 10. A.Uchiyamaand T.Yatabe,Jpn.J. Appl. Phys. 42, 6941(2003). 11. A. Uchiyama, Y. Ono, Y. Ikeda, H. Shuto, and K. Yahata, Poly. J. 44, 995 (2012). 12. K. Kuboyama, T. Kuroda, and T. Ougizawa, Macromol. Symp. 249, 641 (2007). 13. K. Adlem, O. L. Parri, K. Skjonnemand, and D. Wikes, “Mesogenic dimers,” U.S. patent 8,252,389 (August 28, 2012). 14. O. Parri, G. Smith, R. Harding, H.-J. Yoon, I. Gardiner, J. Sargent, and K. Skjonnemand, Proc. SPIE 7956, 79560W (2011). 15. C. A. Guymon, E. N. Hoggan, N. A. Clark, T. P. Rieker, D. M. Walba, and C. N. Bowman, Science 275, 57 (1997). 16. I. Dierking, L. Komitov, S. T. Lagerwall, T. Wittig, and R. Zentel, Liq. Cryst. 26, 1511 (1999). 17. L. D. Sio, P. D’Aquila, E. Brunelli, G. Strangi, D. Bellizzi, G. Passarino, C. Umeton, and R. Bartolino, Langmuir 29, 3398 (2013). 18. L. D. Sio, S. Ferjani, G. Strangi, C. Umeton, and R. Bartolino, J. Phys. Chem. B 117, 1176 (2013). 19. J. L. Janning, Appl. Phys. Lett. 21, 173 (1972). 20. J.-H. Lee and T.-H. Yoon, J. Appl. Phys. 114, 083501 (2013). 21. J.-H. Lee and T.-K. Lim, J. Appl. Phys. 98, 094110 (2005).