1. Introduction

Photopolymers have been found in various engineering applications. For example, they have been used for holographic displays [1], passive and active diffractive optical elements [2, 3], holographic data storage [4], polymer waveguides [5–7] and control of slow-neutron beams [8–10]. In these applications most of photopolymers employ free radical mediated chain-growth polymerization of cross linking (meth)acrylate monomers. This is so because large refractive index changes and rigid polymer network can be obtained under UV/visible illumination. It is known, however, that when van der Waals distances between unreacted monomers are converted to those of their covalent bonding due to polymerization, the resulting microscopic free volume loss causes (bulk) polymerization shrinkage close to or larger than 10%. Such shrinkage distorts formed holographic gratings and refractive index structures, leading to detrimental effects on device performance. Therefore, accurate shrinkage measurements are important for the characterization and development of photopolymer materials.

Aside from conventional density-change measurements of photopolymer materials, the following mechanical and optical methods have also been proposed so far: polymer shrinkage tensometer [11] mechanically measures the evolution of polymerization shrinkage and stress, and it has been used in dental research [12–17]. On the other hand, the noninvasive optical method includes 1) interferometric measurements of temporal/spatial changes in interference fringe patterns [18–21] and 2) holographic measurements of Bragg-angle detuning of either one slanted volume grating [22,23] or multiple weak volume gratings at different slanted angles [24] after recording. The interferometric method can evaluate the time evolution of shrinkage in the out-of-plane (transverse) direction of a photopolymer film. The holographic Bragg-angle de-tuning method can evaluate steady-state shrinkage in the out-of-plane [22–24] and the in-plane (lateral) [23] directions of a recorded photopolymer film. However, the latter method may not be able to sense shrinkage occurred at the early recording stage since a weak volume grating being recorded is continuously refreshed in the low viscous state of photopolymer. We refer such early shrinkage to as dark shrinkage here. For this reason the real-time interferometric measurement is of significant importance to characterize the overall shrinkage properties of photopolymer materials.

Here we report on a simple method of measuring the shrinkage evolution of a photopolymer material during curing by using a white-light (spectral) interferometer with a single broadband beam. In order to examine the validity of the method, we evaluate the shrinkage evolution of a blue-sensitized nanoparticle-polymer composite (NPC) material reported recently [25]. We also compare steady-state shrinkage measured by our new method with that by one of the holographic Bragg-angle detuning methods [24, 25].

2. Method for measuring polymerization shrinkage

Our spectral interferometric measurement system is shown in Fig. 1. The basic idea is the same as that by Toullec et al. [26] who measured the refractive index of dense helium at extremely high pressure. A single broadband probe beam from a fiber-bundle output of a tungsten halogen light source (LS-1-LL, Ocean Optics) was delivered to a sample cell through a lens (f =50 mm). The sample cell consist of two glass substrates on which reflective indium-tin oxide (ITO) electrodes (the center area of 11×11 mm²) were deposited beforehand in order to form a Fabry-Pérot etalon structure supported by a thin photopolymer film that was loaded around the ITO region on one of two glass substrates. Thin spacer films (≈ 10 μm thickness) were also loaded on one of two glass substrates for a guide to form a photopolymer film of appropriate thickness. The transmitted beam that underwent multiple reflection inside the air gap of the sample cell in the Fabry-Pérot etalon structure was detected through a lens (f =50 mm) by a fiber-coupled spectrometer (FICS Spectrograph Model 77443, NEWPORT) equipped with an electrically cooled CCD camera (iDus 420, ANDOR). The spectrograph signal data from the CCD camera were fed into a laptop computer via USB 2.0 connection. The data sampling time was set to be 249 msec, much shorter than the curing time of a photopolymer film used. The onset of curing (curing time t=0) was determined by the appearance of a sharp spectrum at a wavelength (404 nm in our experiment as described below) of a uniformly illuminated curing beam that was slightly scattered into the fiber connected to the spectrometer. In this way the obtained spectrograph at t exhibited a biased periodic pattern (i.e., a spectral interference fringe pattern), as shown by the solid curve in Fig. 2(a). The bias level of the spectral fringe was calibrated in the following way: a spectral signal [the solid curve in Fig. 2(a)] transmitted through the sample cell with the photopolymer film was divided by a pre-registered spectral signal [the dotted curve in Fig. 2(a)]. This pre-registered signal was taken from a transmitted beam through the sample cell without the photopolymer film but with a spacer film of ≈ 100-μm thickness, much thicker than the coherence length of the broadband light source in order to obtain the net spectrum of the broadband light source without any unwanted spectral interference fringe pattern. The resultant calibrated spectrograph, as shown in Fig. 2(b), is used to calculate the order of interference M at a wavelength λ₁ giving one of the maxima of the spectral interference fringe at t, which is given by [26]

 

Fig. 1 Optical setup of white-light interference measurement (above), and a method of making a sample cell (below).

Download Full Size | PDF

 

Fig. 2 Transmitted spectrographs (a) before and (b) after calibration of the background level at t.

Download Full Size | PDF

3. Experimental results and discussions

In the proof-of-principle experiment we employed the blue-sensitized NPC material [25], which consisted of a stoichiometric composition of a commercial secondary thiol monomer [1,4-bis(3-mercaptobutyryloxy) butane (dithiol, Showa Denko K.K.)] and an allyl triazine triene monomer [triallyl-1,3,5-triazine-2,4,6(1H,3H,5H)-trione (TATATO, Aldrich)]. We mixed this thiol-ene formulation with SiO₂ nanoparticles (average size of 13 nm) dissolved in methyl isobutyl ketone to make thiol-ene based NPC mixture. In order to sensitize it in the violet-blue spectral region, we also doped a radical photoinitiator, 2, 4, 6-trimethylbenzoyl-diphenyl-phosphineoxide (DarocurTPO, Ciba) at varied doping concentrations. As shown in Fig. 1, the thiol-ene based NPC mixture was dripped around a patterned ITO electrode deposited on a 10-μm-spacer loaded glass plate and was covered with another ITO-deposited glass plate after drying procedure. The resultant NPC film was made thicker than the spacer film in order not to be clamped by it during and after curing. A highly coherent and expanded blue (404 nm) diode laser beam, also used for our holographic recording and Bragg-angle detuning measurements [25], was employed as an obliquely incident curing beam (see Fig. 1). A curing intensity (I_c) was 5 mW/cm², the optimum holographic recording intensity for the blue-sensitized NPC film [25], unless otherwise stated. It should be noted that either filtered thermal light or an LED can also be used as a curing beam as long as a curing wavelength and intensity are appropriate since high coherence is not required for it.

Figure 3 shows shrinkage evolution as a function of curing time for thiol-ene based NPC films doped with the 1 wt.% photoinitiator at different values for I_c. It can be seen that the rate of shrinkage is higher at higher I_c because of higher polymerization rate at higher I_c. It can also be seen that final values for σ (σ_f) are more or less 4%, independently of I_c. In order to examine the correlation between polymerization and shrinkage evolution, we measured the time growth of conversion (α) of monomer to polymer and that of σ at the photoinitiator concentration of 1 wt.% as shown in Fig. 4, where α_f is a final value for α. Photo-differential scanning calorimetry was used to measure the time-dependent α by means of a commercial photo-calorimeter (Q200, TA Instrument). The detailed measurement method is described in [25]. It can be seen that there is high correlation between the time growth curves of α/α_f and σ/σ_f, which shows the validity of our proposed method. It can also be seen in the inset of Fig. 4 that the polymerization starts slightly earlier than the induction of shrinkage due to the fact that bulk shrinkage is followed by a microscopic free volume loss as a result of polymerization [28].

 

Fig. 3 Shrinkage vs. curing time at different curing intensities for thiol-ene based NPC films doped with the 1 wt.% photoinitiator. Measured initial and final thicknesses (in μm) at t = (0 s, 900 s) by our spectral interferometer were (17.978, 17.228), (19.530, 18.736), (16.530, 15.847), (17.296, 16.574) and (16.899, 16.213) at the curing intensities of 1, 2, 3, 4 and 5 mW/cm², respectively.

Download Full Size | PDF

 

Fig. 4 Normalized conversion α/α_f (dotted curve) and shrinkage σ/σ_f (solid curve) vs. curing time for a thiol-ene based NPC film doped with the 1 wt.% photoinitiator. The inset is the portion of the figure in the early curing time duration.

Download Full Size | PDF

Figure 5(a) shows shrinkage evolution as a function of curing time for thiol-ene based NPC films doped with different concentrations of the photoinitiator. It can be seen that the rate of shrinkage is higher at higher photoinitiator concentrations since the photopolymerization rate increases with an increase in photoinitiator concentration [25]. It can also be seen that σ_f is more or less independent of the photoinitiator concentration, which is summarized in Fig. 5(b) where the measured dependence of σ_f on the photoinitiator concentration (the symbol ○) is plotted. Also, plotted in Fig. 5(b) is a photoinitiator-concentration dependence of σ_f (the symbol •) measured by the holographic Bragg-angle detuning method in which seven weak plane-wave volume gratings were recorded at different slanted angles, followed by uniform post-exposure until their saturation [25]. Bragg-angle changes of recorded plane-wave volume gratings at seven different slanted angles, together with refractive indices of uncured and cured NPC films, were used to find σ_f [24]. It can be seen that the trend of σ_f measured by the spectral interferometric method is quite different from that by the holographic Bragg-angle detuning method.

 

Fig. 5 (a) Shrinkage vs. curing time for thiol-ene based NPC films doped with different concentrations of the photoinitiator. Measured initial and final thicknesses (in μm) at t = (0 s, 900 s) by our spectral interferometer were (17.384, 16.675), (16.306, 15.653), (18.612, 17.901), (18.025, 17.283) and (16.899, 16.213) at the photoinitiator concentrations of 0.1, 0.3, 0.5, 0.7 and 1.0 wt.%, respectively. (b) Dependence of σ_f taken from Fig. 5(a) on concentration of the photoinitiator (○). Measured values for σ_f by the holographic Bragg-angle detuning method (•) and calibrated values for σ_f taken from Fig. 5(a) (□) are also plotted. Solid lines are a guide to the eye.

Download Full Size | PDF

In order to understand this discrepancy, we examined the shrinkage evolution at the early stage of polymerization. Figure 6 shows σ/σ_f and η/η_f vs. curing/recording time for thiol-ene based NPC films at different photoinitiator concentrations. Here η and η_f are the time-dependent diffraction efficiency and its saturated value, respectively, of a plane-wave unslanted volume grating recorded until its saturation, where the uncertainty of η_f for NPC film samples at each photoinitiator concentration is approximately ±10%. Also, σ and σ_f are taken from the spectral interferometric measurement as shown in Fig. 5(a). Each horizontal dotted line corresponds to 10% of η_f until which holographic recording continued, followed by uniform post-exposure in the holographic Bragg-angle detuning measurement as mentioned above. We note that 10% of η_f was experimentally chosen to record weak volume gratings at different slanted angles, which were partially erased after uniform post-exposure, in our holographic Bragg-angle detuning measurement. This recording condition gave detectable diffracted signals to measure the Bragg-angle detuning with high accuracy for shrinkage estimation. Each vertical dotted line corresponds to the value for σ/σ_f at the curing time when 10% of η_f has been reached. In Fig. 5(b) the data (the symbol □) correspond to calibrated values for σ_f when dark shrinkage (i.e., σ at which 10% of η_f has been reached) is subtracted from σ_f obtained by the spectral interferometric method. It can be seen that they are in good agreement with those (the symbol •) obtained by the holographic Bragg-angle detuning method. From this result we can understand the reason why σ_f measured by the holographic Bragg-angle detuning method depends on photoinitiator concentrations as reported in [25]. Namely, the observed photoinitiator-concentration dependence of σ_f [i.e., the symbol • in Fig. 5(b)] can be attributed to the fact that as the photoinitiator concentration decreases, dark shrinkage generally increases (see Fig. 6) so that σ_f measured by the holographic Bragg-angle detuning method is predominantly determined by the remaining shrinkage taken place during the post-exposure period. It follows that although the holographic Bragg-angle detuning method provides useful shrinkage information for holographic applications where uniform post-exposure is made after weak hologram recording (e.g., holographic data storage), it generally gives lower σ_f than that of material’s overall shrinkage. It also suggests that long exposure recording until saturation may give lower σ_f than that by short exposure recording followed by uniform post exposure in the holographic Bragg-angle detuning measurement. Our investigation on this conjecture is underway.

 

Fig. 6 Normalized shrinkage and diffraction efficiency vs. curing/recording time for thiol-ene based NPC films at photoinitiator concentrations of (a) 0.1, (b) 0.3, (c) 0.5, (d) 0.7 and (e) 1.0 wt.%.

Download Full Size | PDF

4. Conclusion

We have proposed a new spectral interferometric method of measuring shrinkage evolution in a photopolymer material. Our proposed method is simple and resistant against environmental disturbances since a single probe beam is employed. It provides the information on the overall shrinkage evolution of a photopolymer material during curing. We have shown its validity by using our developed blue-sensitized NPC material. We have also clarified a discrepancy in steady-state shrinkage between the spectral interferometric and holographic Bragg-angle de-tuning measurements.