1. Introduction

Optical nanocomposite materials [1] are constructed by dispersing an optically functional nano-sized guest material (e.g., nanoparticles) in an optically transparent host matrix. The addition of the guest to the host can yield significant improvement and modification in their optical, mechanical, thermal and electrical properties over bulk materials. Such advantages have led to interesting photonic applications [2]. Since our first demonstration of photopolymerizable titania (TiO₂) nanoparticle-polymer composites [3] as nanocomposite recording materials for holographic data storage, experimental demonstration of volume holographic recording in such recording materials using various inorganic or organic nanoparticles incorporated in photopolymers based on chain-growth free radical polymerization was reported [4–15]. In particular, we showed that the inclusion of nanoparticles in (meth)acrylate photopolymers yielded a substantive reduction in polymerization shrinkage and improved thermal stability as well as an increase in the saturated refractive index modulation (Δn_sat) and the material sensitivity (S) at the same time [16]. Measured values for Δn_sat and S exceeded the required minimum values of 5×10⁻³ and 500 cm/J, respectively, for holographic data storage materials [17]. However, polymerization shrinkage of recorded holograms was of the order of 1% that was still larger than the required criterion of 0.5% [17].

To further reduce shrinkage with photopolymerizable nanoparticle-polymer composites, we recently proposed the use of thiol-ene photopolymerizations that proceed via a step-growth radical addition mechanism [18–20]. The main idea of using the step-growth radical addition polymerization for volume holographic recording stems from the fact that while high-molecular-weight polymer is formed immediately in chain-growth radical polymerization of crosslinking multifunctional (meth)acrylate monomers, it is not obtained until the later stage of the polymerization in the step-growth polymerization process [21, 22]. As a result, gelation takes place late in conversion in the step-growth polymerization process, leading to reduced volume (bulk) shrinkage and stress [23, 24]. Thiol-ene systems use such a step-growth polymerization mechanism based on the radical addition of a thiol to a vinyl (ene) functional group. More specifically, thiol-ene polymerization proceeds by a step-growth radical addition mechanism via sequential propagation of a thiyl radical (RS·) through an ene functional group (R′CH=CH₂) and the subsequent chain transfer of a generated carbon radical (R′C·H-CH₂-SR) to a thiol (RSH), regenerating a thiyl radical, as shown below.

Thus, corosslinked thiol-ene polymerizations proceed very rapidly but will not reach the gel point until high functional group conversions. Other advantages of thiol-ene polymerization include low toxicity, a lack of oxygen inhibition and flexible formulation chemistries based on tailoring thiol-ene monomer chemistry and functionality [25].

Indeed, using organic hyperbranched polymer (HBP) nanoparticle-(thiol-ene)polymer composites with multifunctional mercaptopropionate thiols and allyl ether enes, we showed shrinkage reduction as low as 0.3% with Δn_sat and S as large as 8×10⁻³ and 1014 cm/J, respectively, in the green [19, 20]. The reduced shrinkage was comparable to other low-shrinkage dry photopolymer systems such as those including a high content of inert binder components and using monomers capable of cationic ring-opening polymerization [26]. On the other hand, since the glass transition temperature (T_g) of the cured organic HBP nanoparticle-(thiol-ene)polymer composite was low (∼ −20 °C) due to the formation of flexible thio-ether linkages and the dispersion of organic HBP nanoparticles, the thermal stability of recorded holograms was not as high as that of inorganic nanoparticle-(meth)acrylate polymer composites [16]. To improve the thermal stability with maintaining the advantages of large Δn_sat, high S and low shrinkage, a suitable combination of rigid thiol and ene monomer structures that can allow the uniform dispersion of inorganic nanoparticles needs to be found.

In this paper we report the impact of inorganic silica nanoparticles and an allyl triazine ene monomer [triallyl-1,3,5-triazine-2,4,6(1H,3H,5H)-trione (TATATO, Aldrich)] used for photopolymerizable nanoparticle-(thiol-ene)polymer composites on shrinkage and thermal stability as well as on volume holographic recording. It is shown that our proposed silica nanoparticle-(thiol-ene)polymer composite system provides large Δn_sat, high S and reduced shrinkage characteristics. These values are comparable to those of previously reported organic HBP nanoparticle-(thiol-ene)polymer composites [19, 20]. Furthermore, the use of TATATO results in an increase in T_g and improves the thermal stability of recorded volume holograms, comparable to our previously reported inorganic nanoparticle-(meth)acrylate polymer composites [16]. All these results are attributed to the combination of silica nanoparticles and TATATO that possesses the rigid structure of the triazine group, the high electron density of the double bond and homopolymerization characteristics giving fast thiol-ene polymerization rates, moderately late gel point conversion and increased cross-linking network density [22, 27–29].

2. Experiments

2.1. Sample Preparation and Experimental Methods

We employed a commercial secondary thiol monomer, 1,4-bis(3-mercaptobutyryloxy)butane (dithiol, Showa Denko K.K.). This type of hindered multifunctional thiol monomers has been developed to prevent ambient thermal free-radical polymerization since the shelf-life stability of widely employed primary thiol-ene systems has been a subject of considerable concern [29]. For this reason the secondary dithiol monomer exhibiting high shelf-life stability and low odor was used in our experiment. Also, it showed excellent mixture with the triene monomer (TATATO) and silica nanoparticle (average size of 13 nm) before and after curing and gave better holographic recording performance than other combinations with multifunctional mercaptopropionate thiol monomers. Note that since the average functionality of TATATO is greater than two, a crosslink thiol-ene polymer network can be formed [23,30,31]. The chemical structures of dithiol and TATATO monomers are shown in Figs.1(a) and 1(b), respectively. A stoichiometric composition of thiol and ene functional groups was used in our thiol-ene monomer formulation. Since a refractive index difference between the silica nanoparticle (1.4558 at 546 nm) and the thiol-ene polymer (1.5477 at 546 nm) is close to 0.1, a large Δn_sat is expected. Furthermore, the incorporation of silica nanoparticles would provide better thermal stability of a recorded hologram than the case of using organic HBP nanoparticles since the coefficient of linear thermal expansion (α_L) for inorganics is much smaller than that for organics and the sign of the thermo-optic coefficient (dn/dT) is opposite to that of organics [16]. To obtain efficient step-growth thiol-ene polymerization in the green, we employed 2 wt.% titanocene organometallic complex (Irgacure 784, Ciba) in combination with 2.5 wt.% benzoyl peroxide (BzO₂, Aldrich) to efficiently generate initiating benzoyl oxy radicals that abstract a hydrogen atom from a thiol monomer thereby giving a thiyl radical [32]. All reagents were used as received without further purification.

 

Fig. 1 Chemical structures of thiol-ene monomers used in this study. (a) dithiol, (b) triene TATATO, (c) trithiol, and (d) triallyl ether ene.

Download Full Size | PDF

We mixed the thiol-ene monomer formulation with silica nanoparticles dissolved in methyl isobutyl ketone [4] together with Irgacure784 and BzO₂. Such syrup was cast on a glass plate and was dried at 55 °C for 20 min. in an oven to eliminate the solvent. We refer such a thiol-ene monomer-initiator combination without or with silica nanoparticles to as sample I. To make film samples for optical measurements, we covered the syrup on a 10 μm spacer-loaded glass plate with another glass plate. For comparison we also prepared another nanoparticle-(thiol-ene)polymer composite using a stoichiometric mixture of a mercaptopropionate trithiol, trimethylolpropane tris(3-mercaptopropionate) (Aldrich), and an allyl ether triene, pentaery-thritol allyl ether (Aldrich). This mixture had the refractive index of 1.5319 at 546 nm in the solid phase. The chemical structures of trithiol and triene monomers are shown in Figs.1(c) and 1(d), respectively. The mixture was dispersed with organic HBP nanoparticles having the high-refractive-index of 1.7210 at 546 nm, together with Irgacure784 and BzO₂ at the same doping concentrations as those in sample I. We refer it to as sample II. We note that this nanoparticle-(thiol-ene)polymer composite showed shrinkage reduction as low as 0.3% with the maximized Δn_sat (=8×10⁻³) and S (=1014 cm/J) in the green, as reported previously [19, 20].

Figure 2 shows spectral dependences of absorption coefficients before and after curing for samples I and II with the optimum nanoparticle concentrations of 25 and 35 vol.%, respectively, that maximize Δn_sat as will be shown later. It can be seen that they are more or less 10 and 1 cm⁻¹ at a wavelength of 532 nm before and after curing, respectively, for both samples. This means that their optically available thicknesses are thicker than the minimum media thickness of 500 μm required for holographic data storage [17].

 

Fig. 2 Spectral dependences of absorption coefficients for samples I and II before and after curing under green LED (HDA-TG3, MeCan Imaging Inc.) illumination.

Download Full Size | PDF

To clarify the role of the thiol-ene polymerization kinetics on the holographic recording characteristics, real-time spectroscopic and photocalorimetric studies were conducted. A Fourier transform infrared (FTIR) spectrometer (Nicolet 6700, ThermoFisher Scientific) with a liquid nitrogen cooled mercury cadmium telluride detector was used to measure photo-induced polymerization reactions of thiol and ene functional groups in real time under ambient conditions. Thiol-ene monomers, together with Irgacure 784 and BzO₂, without and with silica nanoparticle dispersion were spin-coated on an aluminum-deposited glass substrate and were placed in a sample chamber where nitrogen gas purging was made for oxygen-free measurements and two KRS-5 windows were attached for incident and reflected infrared probe beams in a reflection mode. Polymerizations were initiated by loosely focused light beam via a lightguide that was connected to a high-pressure mercury arc lamp (SUPER CURE-204S, SAN-EI) through a 532-nm bandpass filter. The curing intensity was set to be 5mW/cm² that was the optimum intensity for holographic recording as described later. Curing was continued until the functional group absorption spectra did not decrease. While thiol functional group conversion was monitored with the S-H stretching absorption peak at 2569 cm⁻¹, allyl (ene) functional group conversion was monitored with the =C-H stretching absorption peak at 3078 cm⁻¹ [28]. Conversions of thiol and ene monomers were respectively calculated by taking the ratio of a change in absorption peak under curing to absorption peak before curing. Note that any background absorption was subtracted from absorption peak beforehand.

We also performed photo-differential scanning calorimetry (photo-DSC) by using a commercial photo-calorimeter (Q200, TA Instrument) to evaluate conversions of thiol-ene mixtures. It was equipped with a refrigerated cooling system (RCS 90, TA Instrument) in order to accurately maintain the isotherm condition at 25 °C. Sample preparation for this measurement was described elsewhere [15]. Time-dependent conversion of thiol-ene mixture α(t) is a ratio of the total number of reacted thiol and ene functional groups at curing time t to that of unreacted thiol and ene functional groups before curing. The enthalpy change due to exothermic reactions over t [ΔH(t) in J mol⁻¹] is proportional to the total number of reacted thiol and ene functional groups at t, so that it can be calculated by the time integration of the measured heat flow (dH/dt in J s⁻¹) from 0 to t with photo-DSC. Therefore, α(t) is obtained by the following relation:

T_gs of cured thiol-ene polymers were measured by photo-DSC and a dynamic mechanical analyzer (Q800, TA Instrument). In the photo-DSC measurement we monitored thermal transitions of approximately 5.5 mg samples in an uncovered aluminum pan by heating/cooling them from −50 to 50 °C at a rate of 10 °C/min under a nitrogen purge. In the dynamic mechanical analysis (DMA) samples of approximately 10×10×0.1 mm³ in size were heated from −50 to 100 °C at a rate of 5 °C/min and at a frequency of 1 Hz.

In holographic measurements we used a conventional two-beam interference setup to write an unslanted transmission grating of 1-μm spacing with two mutually coherent beams of equal intensities from a highly coherent Nd:YVO₄ laser operating at 532 nm. A low intensity He-Ne laser beam operating at 633 nm was employed as a Bragg-matched readout beam to monitor the grating buildup dynamics since the employed initiator system was substantively insensitive in the red. All the beams were s-polarized. We performed measurements at room temperature (25 °C) unless otherwise specified. We measured the diffraction efficiency (η) that was defined as the ratio of the 1st-order diffracted signal to the sum of the 0th- and 1st-order signals. The effective thickness (ℓ) of each sample was estimated from a least-squares curve fit to the Bragg-angle detuning data of the saturated η (η_sat ) with Kogelnik’s formula for an unslanted transmission grating [33]. Then Δn_sat was extracted from η_sat with a help of the estimated value for ℓ and Kogelnik’s formula as given by sin²(πΔn_satℓ/λ cosθ_B), where λ is a readout wavelength in vacuum and θ_B is the Bragg angle. Note that the refractive index modulation (Δn) at 633 nm was converted to that at 532 nm by multiplying the former by a factor being the ratio of Δn_sat at 532 nm to that at 633 nm. Applying the factor to the buildup dynamics of Δn measured at 633 nm, we evaluated S at 532 nm by using the formula given by $(1/I_0\ell )\hspace{0.17em}(\partial \hspace{0.17em}\sqrt{\eta }/\partial t)\hspace{0.17em}{|}_{t=t_{\text{ind}}}$, where I ₀ is an average recording intensity and t _ind is the induction time period. We also evaluated the fractional thickness change σ arising from polymerization shrinkage by means of Dhar et al.’s holographic method [34]. Polymerization shrinkage was also examined by incident angle-dependent transmittance T defined as a ratio of the transmitted power to the incident power with Fresnel correction to two air-glass substrate boundaries of cured film samples under uniform illumination by an expanded and collimated beam from the Nd:YVO₄ laser. Since holographic scattering results in a typical double drop profile in the angular dependence of T when coherently cured film samples suffer from non-negligible shrinkage [15, 35, 36], this measurement qualitatively provides the information on shrinkage for nanoparticle-(thiol-ene)polymer composites relative to that for nanoparticle-[(meth)acrylate]polymer composites. For this purpose thick (∼100 μm) film samples were used in the angular transmittance measurement [36].

2.2. Photopolymerization Characteristics

Conversions of thiol and ene functional groups as a function of curing time were measured by using the real-time FTIR spectrometer. It was observed that although both thiol and ene final conversions reached nearly 100%, the time-dependent conversion and R_p (i.e., the slope of the conversion) of the ene functional group were slightly higher than those of the thiol functional group during the thiol-ene polymerization event. Figure 3(a) shows a parametric plot of these conversions for sample I without silica nanoparticle dispersion, together with the similar plot for sample II without organic HBP nanoparticle dispersion for comparison. It can be seen that for sample I the relative conversion ratio of thiol to ene functional groups is smaller than unity during curing, indicating that the propagation of a TATATO monomer with a thiyl radical is faster than the chain transfer of a carbon radical with an unreacted thiol. It would also indicate that a portion of TATATO monomers homopolymerize each other via reactions with thiyl radicals and/or initiating benzoyl oxy radicals that arise after decomposition of the complex of BzO₂ with the photo-excited isomer of Irgacure 784 [32]. We note that stable allyl pyrrolidino radicals (i.e., photo-cleaved products of the isomer of Irgacure 784 [32, 37]) are considered to act as radical quenchers and not as initiating radicals [38]. On the other hand, the data for sample II show nearly stoichiometric conversion behavior (i.e., the relative conversion ratio is unity and their final conversions are equal to 100%). Such a difference in functional group conversions observed in Fig.3 (a) would reflect on high electron density of the double bond and homopolymerization characteristics of TATATO monomers [27–29].

 

Fig. 3 (a) Parametric dependences of thiol and ene functional group conversions for sample I (●) and sample II (○) without nanoparticle dispersion. (b) Parametric dependences of thiol and ene functional group conversions for sample I dispersed with silica nanoparticles at different concentrations of 0 (●), 10 (○), 20 (□), and 30 (△) vol.%. The grey solid lines shown in Figs.3 (a) and (b) correspond to stoichiometric functional group conversion.

Download Full Size | PDF

Parametric plots of thiol and ene functional group conversions for sample I with silica nanoparticle dispersion are shown in Fig.3(b). It can be seen that thiol conversion tends to take over ene conversion and that these final conversions are not able to reach 100% as the concentration of silica nanoparticles increases. Such a trend is unexpected since thiol monomers cannot homopolymerize: when a thiyl radical abstracts a hydrogen atom from an unreacted thiol, the original thiyl radical reforms a thiol so that there is no net consumption of thiol groups [30]. Namely, thiol functional groups usually undergo repeated chain-transfer reactions that do not have an effect on the overall thiol conversion. We speculate that silica nanoparticles interfere with TATATO monomer’s homopolymerization events and facilitate the reaction of unreacted thiol groups with carbon/benzoyl oxy radicals. We observed that such an effect of the nanoparticle dispersion on thiol-ene conversion is also seen for sample II employing mercaptopropionate thiol and allyl ether ene monomers.

To examine the gel-point conversion x_c of thiol-ene mixtures, we measured dH/dt as a function of curing time for samples I and II by photo-DSC. Using Eqs.(1) and (3), we plotted parametric dependences of α(t) and R_p for samples I and II without nanoparticle dispersion as shown in Fig.4. It can be seen that R_p is much higher for sample I than that for sample II owing to the high reactivity of TATATO monomer. This result suggests an increase in S with sample I as will be shown later. It can also be seen that x_cs for both samples are between 0.45 and 0.50, much larger than ∼ 0.1 for (meth)acrylate monomers [15], indicating that the step-growth polymerization is dominant. According to the Flory-Stockmayer theory for gelation [21, 39], a dependence of x_c on functionalities of thiol and ene groups is given by

 

Fig. 4 Polymerization rates versus conversions for sample I (solid curve) and sample II (dotted curve) without nanoparticle dispersion.

Download Full Size | PDF

 

Fig. 5 Polymerization rate versus conversions for sample I with 0 vol.% (a), 10 vol.% (b), 20 vol.% (c), 30 vol.% (d) silica nanoparticles.

Download Full Size | PDF

The photocalorimetric measurement showed that the cured sample I without silica nanoparticle dispersion had T_g of −5 °C which was higher than T_g of −20 °C for the cured sample II without silica nanoparticle dispersion. This result is attributed to the fact that the while the formed thio-ether linkages due to the allyl ether ene monomer are flexible in sample II, the rigid ring structure of TATATO monomer in sample I limits the thermal motion of the formed polymer. It was also found that T_g decreased from −5 to −15 °C with an increase in silica nanoparticle concentration from 0 to 30 vol.%. This trend may be explained by a slight increase in unreacted thiol and ene functional groups remaining in the formed polymer as the silica nanoparticle concentration increases (see Figs. 3 and 5 showing a decrease in final conversion with increasing silica nanoparticles). We estimate the fraction of completely unreacted thiol (ene) monomer to be (1 −0.94)² = 0.36% [(1 – 0.89)³ = 0.13%] at 30 vol.% silica nanoparticle dispersion [23]. On the other hand, it was found from the DMA measurement that sample I had T_g of ∼ 0 °C, almost independently of the silica nanoparticle concentration. The measured values for T_g were higher than those in the photocalorimetric analysis, as similar to other studies [40, 41]. This trend is considered to result from the faster scanning frequency for DMA versus DSC [41].

2.3. Holographic Recording Characteristics

Figure 6 shows the buildup dynamics of Δn at different I ₀s for sample I with 25 vol.% silica nanoparticle dispersion that is the optimum value for maximizing Δn_sat as shown below. It can be seen that, as similar to the case of sample II reported previously [19], Δn_sat is maximized near I ₀=5 mW/cm². It can also be observed that the induction time period appears at I ₀s lower than ∼ 10 mW/cm². We speculate that since the oxygen scavenging process generates a thiyl radical by reaction with a thiol [32], allyl pyrrolidino radicals play a role in quenching initiating benzoyl oxy radicals so that the induction event takes place. Hereafter, holographic gratings discussed were recorded at I ₀=5 mW/cm² unless otherwise specified.

 

Fig. 6 Buildup dynamics of Δn at I ₀= 1 (a), 5 (b), 10 (c), 50 (d) and 100 (e) mW/cm² for sample I with 25 vol.% silica nanoparticle dispersion.

Download Full Size | PDF

Figure 7 shows a measured angular selectivity of η_sat for sample I with 25 vol.% silica nanoparticle dispersion. Excellent curve fit indicates that the recorded hologram is thick and the plane wave hologram is uniformly recorded along the thickness direction. Figure 8 illustrates grating-spacing dependence of Δn_sat for sample I with 25 vol.% silica nanoparticle dispersion. It can be seen that there exists a cut-off grating spacing near 0.1 μm from which Δn_sat increases with increasing a grating spacing. Such a grating-spacing dependence was also observed in other nanoparticle-[(meth)acrylate]polymer composites [3, 4, 9]. This peculiar trend may be attributed to the mutual diffusion and phase separation of reactive monomers and unreacted secondary species (nanoparticles in our case) in multi-component photopolymer systems during holographic exposure [42, 43].

 

Fig. 7 Measured angular selectivity of η_sat from a recorded hologram at a 1 μm grating spacing for sample I with 25 vol.% silica nanoparticle dispersion. Bragg angle detuning was evaluated in a glass substrate. The solid red curve corresponds to the least-squares-fit of the data to Kogelnik’s formula for an unslanted transmission grating with Δn_sat and ℓ as fitting parameters. Extracted values for Δn_sat and ℓ were 1.0×10⁻² and 14.3 μm, respectively.

Download Full Size | PDF

 

Fig. 8 Grating-spacing dependence of Δn_sat for sample I with 25 vol.% silica nanoparticle dispersion.

Download Full Size | PDF

Figure 9 shows dependences of Δn_sat [Fig.9(a)] and S [Fig.9(b)] on nanoparticle concentration for sample I. The results for sample II are also plotted for comparison. It can be seen from Fig.9(a) that there exists the optimum nanoparticle concentration of 25 vol.% maximizing Δn_sat for sample I, which is lower than 35 vol.% for sample II. Although the refractive index difference between the formed polymer and nanoparticle for sample I is smaller by approximately 0.1 than that for sample II, sample I has higher Δn_sat at the optimum nanoparticle concentration. This result may be attributed to sample I’s higher transparency and higher crosslinking networks that facilitate efficient mutual diffusion and phase separation of reactive monomers and nanoparticles during holographic exposure [42,43]. The observed concentration dependence of Δn_sat as also observed in other nanoparticle-polymer composites can be explained as follows: when no nanoparticle is dispersed, no steady-state refractive index modulation is created since unreacted monomer is eventually polymerized in the dark region of the intensity-interference fringe pattern. As the doped nanoparticle concentration increases, Δn_sat increases due to the mutual diffusion of monomer molecules and nanoparticles during holographic exposure. However, too much dispersion of nanoparticles causes a decrease in the mutually diffusing amount of monomer molecules and nanoparticles, leading to a decrease in Δn_sat [3, 4, 8, 9]. It can be seen from Fig.9(b) that S is maximized at the same optimum nanoparticle concentration as that for Δn_sat. This is so because S is approximately given by πΔn_sat /(λI ₀ τ) for single exponential grating buildup [13], where λ is the recording wavelength in vacuum and τ is a buildup time constant. Since sample I has higher R_p and larger Δn_sat than those of sample II, it has higher S. We note that values for Δn_sat and S given at and near the optimum nanoparticle concentration are larger than the required minimum values of 5×10⁻³ and 500 cm/J [17], respectively, for holographic data storage.

 

Fig. 9 Nanoparticle concentration versus (a)Δn_sat and (b)S for samples I (●) and II (○).

Download Full Size | PDF

Figure 10 shows a dependence of σ on nanoparticle concentration for samples I and II, together with the result for a silica nanoparticle-[(meth)acrylate]polymer composite sample [4] for comparison. It can clearly be seen that σ exponentially decreases with increasing nanoparticle concentration for all the samples. The reason for such a decrease in shrinkage is attributed to an increase in rigidity of the sample material resulting from the dispersion of inorganic nanoparticles. It can also be seen that σ s for samples I and II are much lower than that for the (meth)acrylate-based sample. It is just below 0.5% for sample I at the optimum nanoparticle concentration (25 vol.%). A factor of 10 reduction in σ is evident as compared with the (meth)acrylate-based sample and other nanoparticle-[(meth)acrylate]polymer composites [3, 4, 7–9, 13]. The observed trend in shrinkage reduction with thiol-ene systems can be explained by our result shown in Fig.4: samples I and II based on step-growth polymerization have x_cs that are 0.45 – 0.50, i.e., gelation takes place much later in conversion than the case of (meth)acrylate monomers as discussed earlier. Indeed, there is a strong correlation between x_c and shrinkage in thiol-ene based nanoparticle-polymer composites [44]. We note that σs for samples I and II at and above the optimum nanoparticle concentration are below the acceptable maximum shrinkage of 0.5% [17] for holographic data storage.

 

Fig. 10 Nanoparticle concentration vs. σ for samples I (●), II (○) and a (meth)acrylate-based sample (□).

Download Full Size | PDF

Figure 11 shows an incident-angle dependence of T for a thick film sample I with 25 vol.% silica nanoparticle dispersion. The result for a (meth)acrylate-based thick film sample with 35 vol.% zirconia nanoparticle dispersion [15] is also plotted for comparison. These thick film samples having thicknesses of approximately 100 μm were cured beforehand under homogeneous and normal incidence exposure by a well-collimated coherent beam at 532 nm. They were probed by the same beam. It can be seen that both film samples suffer from scattering losses particularly near the normal incidence. Such a trend can be explained by holographic scattering due to noise grating formation during recording [36, 45]. Furthermore, T for the (meth)acrylate-based thick film sample exhibits a typical double drop profile in the angular dependence. This happens when the deformation of Ewald sphere consisting of Bragg-matched and self-amplified scattered light from noise gratings takes place as a result of polymerization shrinkage of a holographic recording material [15, 35, 36, 46]. On the other hand, we note that such a characteristic profile is not noticeable for the thick film sample I because of the reduced polymerization shrinkage as seen in Fig.10.

 

Fig. 11 Incident-angle dependences of transmittance T at a wavelength of 532 nm for a thick film sample I with 25 vol.% silica nanoparticles (solid curve) and a (meth)acrylate-based thick film sample with 35 vol.% zirconia nanoparticles (dotted curve) after coherent uniform exposure.

Download Full Size | PDF

We further examined temperature-induced optical and mechanical distortions of plane-wave volume holograms recorded in samples I and II to examine environmental thermal stability. Figure 12 shows nanoparticle concentration dependences of thermo-optic coefficient dn/dT for uniformly cured samples I and II at 25 °C and at a wavelength of 546 nm. For comparison, the data for a silica (34 vol.%) nanoparticle-[(meth)acrylate]polymer composite film [16] is also plotted. It can be seen that while |dn/dT| for sample I is a decreasing function of silica nanoparticle concentration, sample II is an increasing function of organic HBP nanoparticle concentration. This different trend can be explained by the fact that while the refractive indices of inorganic materials are mainly determined by the temperature-dependent polarizability and thus dn/dT is positive, dn/dT is negative for organic materials including organic HBP nanoparticles because of their strong dependence of the refractive index on temperature-dependent volume change [47]. Since sample I contains inorganic silica nanoparticles, |dn/dT| is a decreasing function of silica nanoparticle concentration. The silica nanoparticle-[(meth)acrylate]polymer composite sample has the smallest |dn/dT| among the other thiol-ene-based samples owing to high crosslinking networks of the formed (meth)acrylate polymer.

 

Fig. 12 Thermo-optic coefficients dn/dT at 25 °C and at a wavelength of 546 nm as a function of nanoparticle concentration for uniformly cured film samples I (●), II (○) and a silica nanoparticle-[(meth)acrylate]polymer composite film sample (□).

Download Full Size | PDF

We also estimated α_L from the measurement of temperature-dependent Bragg-angle detuning by means of Dhar et al.’s holographic method [34]. The detailed experimental procedure is described elsewhere [16]. Figure 13 shows out-of-plane α_Ls as a function of nanoparticle concentration. It can be seen that α_L for sample II has an opposite concentration dependence to other samples incorporating inorganic silica nanoparticles. This is so because the high value of α_L of polymers is caused by the low energy barrier for the chain conformation to be changed [48]. On the other hand, the addition of silica nanoparticles drastically reduces α_L. This happens because changes in chain conformation of a formed polymer are constrained at boundaries of silica nanoparticles that have very large surface areas relative to their volumes. Once again, the best performance with the silica nanoparticle-[(meth)acrylate]polymer composite sample is seen owing to the dispersion of silica nanoparticles and the high crosslinking networks in the formed (meth)acrylate polymer. Decreases in |dn/dT| and α_L lead to a reduction in temperature-induced optical distortions of volume holograms recorded in nanoparticle-polymer composite films since thermal changes in refractive index and volume additively alter the Bragg-angle selectivity [49].

 

Fig. 13 Linear coefficients of thermal expansion α_L as a function of nanoparticle concentration for uniformly cured film samples I (●), II (○) and a silica nanoparticle-[(meth)acrylate]polymer composite film sample (□).

Download Full Size | PDF

Figure 14 shows a temperature dependence of out-of-plane thickness change for samples I with 25 vol.% silica nanoparticle dispersion, sample II with 35 vol.% organic HBP nanoparticle dispersion and a silica (34 vol.%) nanoparticle-[ (meth)acrylate]polymer composite sample. Note that while thickness changes at 25 °C correspond to polymerization shrinkage, those at other temperatures are caused by temperature deviations of recorded holograms from 25 °C. Substantive improvement in dimensional stability over the measured temperature range is evident for sample I as compared with sample II. Although the temperature stability of the silica nanoparticle-[(meth)acrylate]polymer composite sample is better than that of sample I, its absolute thickness changes are out of the acceptable range (±0.5 %) for holographic data storage. On the other hand, sample I allows relatively large temperature changes between 25 and 55 °C within the acceptable range.

 

Fig. 14 Temperature dependence of out-of-plane thickness change measured in percent for sample I (●), sample II (○) and a silica nanoparticle-[(meth)acrylate]polymer composite sample (□). The colored band corresponds to the thickness change within ±0.5 %. Note that the thermally induced small thickness change results in a change in the diffraction efficiency as twice as the percent thickness change for each hologram having low diffraction efficiency in hologram multiplexing.

Download Full Size | PDF

Finally, we present a result of holographic recording of a two dimentional (2D) digital data page pattern (384×384 binary pixels) generated by a liquid crystal spatial light modulator (LC2002, HOLOEYE) with sample I with 25 vol.% silica nanoparticle dispersion. Fourier transform holography setup was employed to record the 2D pattern. Figure 15 illustrates a input image through an optical system [Fig. 15(a)] and the corresponding reconstructed image [Fig. 15(b)]. The output image with good fidelity is seen. A detailed study on the hologram multiplexing capability of our nanoparticle-(thiol-ene)polymer composites is underway.

 

Fig. 15 Holographic recording of a 2D digital data page pattern using sample I with 25 vol.% silica nanoparticle dispersion: (a) an input image through the optical system and (b) a reconstructed image.

Download Full Size | PDF

3. Conclusion

We have described high performance holographic nanoparticle-polymer composites using the thiol-ene chemistry of photopolymerization. We have found that the stoichiometric mixture of secondary dithiol and triallyl TATATO ene monomers partially exhibits homopolymerization behavior during curing as a result of the high electron density of TATATO monomer. It has been shown that the mixture gives the increased peak value for R_p by more than a factor of 3 as compared with that of the mixture of primary trithiol and triallyl ether ene monomers. We have also found that both thiol-ene mixtures have the gelation point at later conversion. Such an occurrence can accelerate the mutual diffusion and phase separation of monomer molecules and nanoparticles during recording, resulting in an increase in Δn_sat and S as observed. The polymerization shrinkage can also be suppressed due to an increase in gelation point. Obtained values for Δn_sat, S and shrinkage meet the media requirement for holographic data storage. We have also investigated thermal distortions of recorded volume holograms. We have shown that thermal thickness changes are suppressed owing to decreases in both |dn/dT| and the out-of-plane α_L by the use of inorganic silica nanoparticles and TATATO monomer. Such an improvement, together with increases in Δn_sat and S and with a substantive reduction in shrinkage, suggests the usefulness of inorganic nanoparticle-polymer composites using step-growth thiol-ene systems for holographic data storage and other photonic applications.