1. Introduction

Epoxies are widely used in different applications in everyday life. The properties of epoxies allow us to fabricate durable structures with high optical quality. Luminescent epoxies have been recently studied as coatings for light emitting diodes (LEDs) [1–5], as sensors [6, 7], in waveguides [8], and in solar concentrator applications [9–11]. Luminescent solar concentrators (LSCs) were discovered in 1970s to lower the cost of solar cells by reducing the area of solar cell [12, 13]. These concentrators consist of a polymer containing luminescent dye. The matrix material used in solar concentrators should have high transmittance in visible region and the dye should have low self-absorbance and high quantum efficiency. Organic dyes are known to be sensitive to UV irradiation and can easily undergo degradation. Sealing of dye-sensitized solar cell devices (DSSCs) with photo-curable fluorinated polymers [14] and fluorine-containing UV-curable adhesives [15] has recently been shown to increase the long-term stabilities of DSSC-systems.

In other applications, such as LED coatings, also the extraction of light from the structures plays an important role besides the material properties. Different types of surface patterning can be used in order to avoid total internal reflection inside the LED structures and thus to increase the extraction efficiency [16–18]. Micro- or nanopatterning can also be used for other purposes, including beam shaping [19] and antireflective, hydrophobic, or self-cleaning surfaces [20,21]. Surface patterns can be implemented on epoxy materials with direct lithographic methods as has been demonstrated, e.g., for an epoxy-based photoresist containing luminescent nanocrystals [22]. More recently, the transfer of microstructures on pure epoxy surface using a simple stamping process has been studied [23]. However, such approach has not been previously used for patterning of luminescent optical epoxies.

In this work we demonstrate the transfer of micropatterns on the surface of epoxy samples containing different luminescent dyes by imprinting and study the luminescence properties of the patterned and unpatterned samples. We have used epoxy resin to protect the dyes from oxidation and degradation. A similar approach has been used in the case of DSSCs to protect solar cell components from photochemical degradation [14,15]. Luminescent dyes were separately mixed with Bisphenol A (BPA) epoxy resin which was cured with Isophorone diamine hardener. In addition, an epoxy sample with two different luminescent dyes was fabricated in order to test color mixing. To test the replication of microstructures, two linear gratings with different triangular profiles were imprinted on the surface of epoxy samples during the curing using a hot embossed cyclo-olefin polymer mold. The absorbance spectra of the unpatterned epoxy samples and the luminescence intensities of all fabricated luminescent epoxy samples were measured using a custom-made bispectrometer. Our replication method offers a simple and low-cost alternative for patterning luminescent epoxies for different applications. Part of the results presented in this paper has been previously presented in a conference paper [24].

2. Fabrication

Bisphenol A (Bisphenol A diglycidyl ether, tech. 80%, Alfa-aesar) was used as the epoxy resin. The epoxy resin was stirred at 80°C for 60 min prior curing in order to reduce the number of air bubbles. When fabricating the luminescent samples 0.016-0.178 wt-% of the luminescent dye was added to the epoxy resin (Table 1) and the mixture was stirred. 22.7 wt-% of Isophorone diamine (IPDA, Aldrich) as a curing agent was added to the mixture of epoxy resin and luminescent dye, and the mixture was further stirred for 5 min. The epoxy was placed on a silicone mold and was cured at 80 °C for 120 min. The masses of epoxy resin, mass-% of the luminescent materials compared to the epoxy resin and the amounts of hardener are presented in Table 1. Images of epoxy samples illuminated with a standard D65 lamp simulating daylight and a UV lamp with peak wavelength at 365 nm are shown in Fig. 1. It should be noted that some of the samples are excited at shorter wavelengths than those included in the used UV source, so their luminescence is not shown in these photographs.

Table 1. Used mass-% (compared to epoxy resin) of the luminescent material, the excitation wavelengths, and the best emission wavelengths of the epoxy samples.

View Table

 

Fig. 1 To demonstrate the luminescence effect and the colors of the fabricated luminescent epoxy samples were illuminated with a D65 lamp (upper images) and a UV lamp (lower images). In the figure: pure epoxy (a), Erythrosin B (b), Fluorescein (c), 2,5-diphenyloxazole (d), Rhodamine B (e) and trans-Stilbene (f) mixed with epoxy.

Download Full Size | PDF

Two different patterns were fabricated to be imprinted on the epoxy surface: a triangular crest grating with period of 2 µm and height of 1.5 µm which could decrease the reflection at the surface of the samples, and a blazed grating with period of 10 µm and height of 3 µm which is similar to the structures used by A. Bay et al. for extracting more light from their samples [17,18]. These test structures allow us to study the replication of patterns with different shapes and dimensions. In particular, they show whether it is possible to replicate deep structures with our method, while the replication fidelity of the sharp features of the patterns can also be used as an indication of the nanopatterning potential.

The triangular grating was machined with Moore Nanotechsys 350FG ultraprecision diamond freeform pattern generator and the blazed grating was fabricated using electron beam lithography and silicon wet etching. From both of the patterns a nickel plate was made using electroplating. The patterns were copied from nickel plates to COP (Cyclo Olefin Polymer, ZF14-188, Zeonex) films by hot embossing. The patterns were transferred to the epoxy surface as follows: Patterned COP films were carefully placed on top of the liquid epoxy in the silicone mold. After the epoxy was hardened by curing at 80 °C and removed from the silicone mold, the COP films were carefully removed. Figure 2 shows scanning electron microscope (SEM) images of cross-sections of the fabricated gratings and the used measurement geometry. The final thickness of all fabricated epoxy samples was approximately 3.5 mm.

 

Fig. 2 SEM images of micro patterns imprinted on epoxy. On left triangular crest grating and on right a blazed grating. The white arrows show incoming light and measurement angle of 45 degrees.

Download Full Size | PDF

3. Measurements

The transmittance and absorbance spectra (Fig. 3) of the unpatterned samples were first measured with a UV/Vis/NIR spectrometer (Perkin Elmer Lambda 900). The excitation-emission matrices (EEMs) of all samples were measured with a custom-made bispectrometer [25–29]. The light from a 450 W Xenon lamp (Oriel M-66923 housing, Newport Corporation, Irvine, California and Osram XBO 450 W bulb, Osram AG, Winterthur, Switzerland) was directed to the sample through a Czerny-Turner monochromator (DTMc300, Bentham Instruments Ltd, Berkshire, United Kingdom). The used peak width and sampling interval of the excitation light were both 5 nm. The emitted light was collected with a lens to an optical fiber attached to a spectrograph detector (PMA-12, Hamamatsu Photonics K.K., Japan). The spectrograph operates at the wavelength range 200–950 nm with 0.72–0.76 nm spectral sampling. In the used measurement geometry angle of incidence was 0° and detection angle was 45° (see Fig. 2.). The excitation peak was removed from the measured EEM.

 

Fig. 3 Measured absorbances (left) and transmittances (right) of pure epoxy and different dye- epoxy mixtures.

Download Full Size | PDF

4. Results & discussion

The absorbance spectra of the samples measured with UV/Vis/NIR spectrometer are presented in Fig. 3(a). Transmission measurements presented in Fig. 3(b) show high transmittance (appr. 90%) of the epoxy matrix in visible region (350–800 nm), which is a requirement for a good luminescent solar concentrator. The wavelength ranges of the absorption bands depend on the dye materials mixed in the epoxy, while the absorption of UV radiation is due to the epoxy itself which is excited in UV and emits near visible (310 nm).

The transfer of gratings into the epoxy surface was successful. The micropatterns were copied to the epoxy surface well, and the method is simple and reproducible. The SEM images of the cross-sections of the cleaved gratings in the epoxy surface are presented in Fig. 2. Since even the small details of the micropatterns, such as sharp grooves and edges of the triangular profile, were transferred to the epoxy surface properly, it suggests that the same technique could be used also for nanopatterning. Further, nanopatterning of various epoxy samples has already been demonstrated using slightly different replication methods [30,31]. Therefore fabrication of anti-reflection, hydrophobic and self-cleaning surface structures directly on the epoxy surface (e.g. directly on top of solar cell) is achievable.

The measured luminescence intensities of the unpatterned and patterned luminescent epoxy samples at best excitation wavelengths are presented in Fig. 4(a)–4(f). The emission wavelengths of maximum luminescence intensities and the corresponding best excitation wavelengths are presented in Table 1. The used epoxy resin has also luminescent properties [32] which were measured, see Fig. 4(a). The luminescence emission spectra for pure epoxy, as well as for trans-Stilbene and 2,5-diphenyloxazole, Figs. 4(b) and 4(c), have lots of disturbances due to the short exposure time (20 ms) that was used in order to avoid saturation of the signal for more strongly luminescent samples. However, the exposure time and the other measurement parameters were kept constant throughout all measurements so the luminescence intensity values are comparable among all results.

 

Fig. 4 Measured luminescence intensities of samples with unpatterned and patterned (triangular or blazed grating) for pure epoxy (a), Erythrosin B (b), Fluorescein (c), 2,5-diphenyloxazole (d), Rhodamine B (e) and trans-Stilbene (f) mixed with epoxy. For each dye the excitation wavelength is given.

Download Full Size | PDF

The luminescence was enhanced by the gratings in most of the samples. It should be noted that the parameters of our test grating structures are not specifically optimized to enhance the extraction of emitted light, since the optimal parameters would have been different for each sample, depending on the used dye materials and the emission wavelengths. Therefore the increase in the luminescence signal caused by the gratings is relatively low in most of the samples. However, as has been demonstrated in [17,18] on which our test structure with blazed profile was based, when the period and height of the grating are suitably chosen regarding to the refractive index of the samples and the wavelength of the emitted light, higher increases in the light extraction are possible.

In all epoxy-dye samples, except 2,5-diphenyloxazole, the triangular grating enhanced the luminescence more than the blazed grating, which is most likely due to the difference in the geometry of the gratings. The triangular grating profile can be considered roughly as a large-scale and one-dimensional version of antireflection surfaces based on nanopyramids [20]. Based on our rigorous numerical simulations using the Fourier modal method [19], the triangular profile couples more incident light into epoxy than the blazed profile, so the triangular grating reduces the reflection at the surface which increases the amount of exciting light in the samples. This effect could also be further improved by using nano-scale structures and optimizing the grating parameters for each sample separately.

The luminescence intensities of the unpatterned epoxy sample containing two different dyes, 2,5-diphenyloxazole and Rhodamine B, are shown in Fig. 5 at two excitation wavelengths. Two separate emission peaks are detected, one for 2,5-diphenyloxazole and one for Rhodamine B, with same excitation and emission wavelengths as in samples containing only one dye. The use of two or more dyes improves the UV to visible conversion and, in addition, allows color mixing in emission.

 

Fig. 5 Measured luminescence for epoxy containing both 2,5-diphenyloxazole and Rhodamine B excited at wavelengths 335 nm and 560 nm (left) and the sample illuminated with D65 lamp and UV lamp (right).

Download Full Size | PDF

5. Conclusions

We have presented an efficient and straightforward way to produce optical epoxy samples containing different commercial luminescent dyes and demonstrated the transfer of micropatterns on the epoxy surface using an effortless and reproducible imprinting method. Based on our spectral measurements, the epoxy shows autoluminescence excited by UV light, which could enable the use of UV radiation from sun in solar concentrators. Furthermore, the luminescence from the dyes converts UV radiation to visible (trans-Stilbene, 2,5-diphenyloxazole), as well as gives the samples bright color. Through color mixing, there are considerable amount of possible colors available to be used, for example, on top of solar cells.

In our fabrication experiment, the micropatterns were copied well in the epoxy surface, which suggests that nanopatterning of luminescent epoxies could also be possible using the same method. This would allow simple fabrication of anti-reflective, hydrophobic and self-cleaning structures, for example. The luminescence intensities of both patterned and unpatterned epoxy samples were measured with bispectrometer to observe the luminescent properties of the luminescent dyes mixed with epoxy. The luminescence was increased by the micropatterns in most of the samples. However, to obtain better enhancement of the luminescence and light extraction, which would be useful, e.g., in LED applications, the surface patterns should be optimized individually for each sample taking into account the emission wavelength and the refractive index of the dye-epoxy mix. Same replication methods could also be used for patterning other polymer materials containing luminescent dyes, for which similar results can be expected.

In addition to the good optical properties of epoxy, it is also very durable material. Thus the epoxy matrix could also protect the luminescent material from, e.g., oxidation and wear. The materials and methods used in this work can be further applied for developing easily available, strong, and low-cost alternatives for many applications, such as large-scale solar concentrators or color coatings.