1. Introduction

Two-photon fluorescence (or photoluminescence) (TPF or TPP) provides high sensitivity detection methods [1–3] for biochemical sensing applications. For organic molecules, two-photon excitation wavelengths are generally in the red or near-infrared range and the emission is in the blue-green visible wavelength range. This relatively large spectral difference can dramatically increase the detection sensitivity by effectively suppressing the background noise, compared to the large background noise caused from autofluorescence in one-photon fluorescence (OPF) detection. Furthermore, two-photon excitation wavelengths are more appropriate for many biological and fluorescent samples, since they easily undergo photobleaching when irradiated by UV or visible light [4]. Although TPP has many advantages than OPF, conventional TPP requires high photon density in focused laser pulses [2,3], which make it less accessible technically. Previously, a promising method, namely optical resonant device [5], was introduced to obtain the needed intensity without tightly focusing of a laser beam.

Resonant waveguide grating (RWG) [5–9] has been used to successfully enhance the fluorescence signal excited by one-photon [7,9] or two-photon [5,10,11] absorption process. This structure typically comprises a high refractive-index grating-waveguide layer and a low refractive-index supporting layer, and can generate very sharp reflection and near zero transmission anomalies that signature the guided mode resonance (GMR) [9]. Thanks to GMR in the RWG, enhancement in excitation or extraction of fluorescence signal can be achieved by aligning either the excitation wavelength or the emission fluorescent spectrum with GMR. When resonant mode coincides with the wavelength of the excitation laser, the enhanced excitation in the RWG can produce strong local field at the grating layer [12–14]. If fluorescent dye tagged biomolecules are in direct contact with the grating, the fluorescence signal can be improved with this strong local field, thus the detection sensitivity. Additionally, when the emission wavelength of a fluorescent dye overlaps with GMR mode, the associated high reflection efficiency [12] will substantially enhance the fluorescence emission. Soria et al. have demonstrated for the first time to enhance TPP by a factor of nearly 330 in the double grating waveguide structure [5]. However, their results focused mainly on the effect of enhancement of the excitation light. Recently, Cunningham et al. [12,15] have revealed that OPF can be multiplicatively enhanced when combining enhanced excitation and enhanced extraction effects. It is expected that such multiplicative effect can be used to further enhance TPP signal.

In this work, we have designed and fabricated one-dimensional (1D) sinusoidal RWG devices with a layer of fluorescent polymer (polyfluorene, PFO) on top which mimics the detection target for bio-sensing applications. A simple and low-cost two-beam interference technique [16,17] and electron-beam deposition method were employed to fabricate the RWG on the scale of mm’s. The numerical calculation method based on rigorous coupled-wave analysis (RCWA) [18] was used to simulate the optical properties of the one-dimensional sinusoidal PFO RWG. We demonstrate multiplicative enhancement in TPP by combining enhanced excitation and enhanced extraction effects.

2. Experiments and numerical simulation

2.1 RWG design and sample preparation

Figure 1(a) shows the design of the PFO RWG structure. It consists of, from the top to the bottom, a PFO polymer layer, a 1D waveguide-grating layer (TiO₂), a cladding layer (SU8), and a glass substrate. The TiO₂ layer was designed to be 55 nm thick (T_TiO2) with periodicity (Λ) of 385 nm such that a TE GMR mode at normal incidence would occur at around 650 nm wavelength. The procedure of sample fabrications follows. A flat layer of SU8 photoresist with a thickness (T_SU8) of 1 µm ± 0.1 µm was first spin-coated on the top of the glass substrate and photopolymerized. A second layer of SU8 film was then spin-coated on top of the first SU8 layer, on which grating structure were produced with a two-beam interference technique [16,17]. A TiO₂ layer with a thickness of 60 nm ± 5 nm was deposited on top of SU8 grating with electron-beam deposition technique. The semiconducting PFO polymer was chosen as the fluorescent emission material. The PFO polymer was first dissolved in toluene then spin-coated on the top of the TiO₂ layer. The thickness (T_PFO) of the PFO polymer was measured to be 260 nm ± 20 nm. Figure 1(b) shows the scanning electron microscopy (SEM) image of a cross-sectional view of the fabricated PFO RWG structure. The period (Λ) and the groove depth (d) of the grating were determined to be 390 nm and 90 nm ± 5 nm by SEM and AFM measurements, respectively. The overall area of the PFO RWG is 6 × 6 mm². According to ellipsometric measurement, the refractive indices of PFO, SU8 and TiO₂ are 1.60, 1.56 and 2.20 at 0.633 μm wavelength, respectively. The refractive indices of these materials at different wavelengths were also obtained from this measurement, which were used in the RCWA simulation. Figure 1(c) plots the absorption and TPP emission spectra measured from a PFO polymer thin film. The absorption mainly occurs between 310 and 420 nm with peak position at 385 nm, and similarly TPP emission is between 500 nm- 650 nm with peak position at 545 nm.

 

Fig. 1 (a) Schematic of a 1D PFO RWG. (b) SEM image of PFO RWG. From top to bottom are: PFO polymer layer, 1D sinusoidal-wave waveguide-grating layer (TiO₂), SU8 polymer layer and glass substrate. The lattice constant (Λ) and the groove depth (d) of grating are 390 nm and 90 nm ± 5 nm, respectively. (c) Absorption and TPP spectra of PFO thin film.

Download Full Size | PDF

2.2 Experimental setup

Figure 2 shows the experimental setup for measuring TPP intensity generated from the sample. An optical parametric oscillator (OPO) nanosecond pulse laser, with tunable wavelength range from 0.79 to 0.84 μm, 5 ns pulse width, and 10 Hz repetition rate, was used as the excitation beam. The excitation power was adjusted with the combination of a half-wave plate and a polarizer. A collimation system that contains two lenses (L₁ and L₂) was used to collimate the incident beam with a beam size of 2 mm diameter. The average intensity of the excitation beam is fixed at 32 mW/cm² for all measurements, and no damage of the sample was observed. An aperture of 4 mm diameter was placed at a distance about 6 cm from the sample to limit the solid angle for collecting TPP intensities to about 0.003 str (~1.9^ο). The TPP signal is then focused by a regular lens (L₃) toward a photomultiplier (PMT) detector, which is connected to a boxcar integrator system. Before the signal reached the PMT, three IR filters were used to block the excitation beam, and an interference filter with central wavelength at 550 nm ± 10 nm was used to select TPP signal at the peak wavelength. The transmission spectra of the RWG sample were measured with a halogen white light source and a grating spectrometer (Andor Shamrock SR-500i). For the measurement of the incident angle (θ_i) dependence of the transmission spectrum, the sample was rotated with respect to the incident beam. The incident plane was arranged to lie on the X-Z plane all the time, where the X-axis is the direction of the grating vector of the RWG sample and the Z-axis is the normal direction of the top surface of the RWG sample, as shown in Fig. 1(a). The precision of the angular position of the sample was 0.5 degrees. For the measurement of the dependence of the TPP signal on the collection angle (θ_c), the signal collection arm orientation was adjusted with respect to the sample, and the precision of the signal collection angle was within 0.5 degrees.

 

Fig. 2 TPP measurement setup. λ/2: half-wave plate, P: polarizer, L₁-L₃: lenses, θ_i: incident angle of excitation beam, θ_c: collected angle of TPP signal, and Filters: three IR filters and one 550 nm interference filter.

Download Full Size | PDF

2.3 Numerical simulation

The simulation tool based on RCWA method [18–21] was employed to design and analyze optical properties of the PFO RWG structure. In the RCWA simulation the grating structure is sliced into layers so that each layer is homogenous in the propagation direction (Z-axis). In each layer, electromagnetic field and permittivity are expanded to a Fourier series in terms of spatial harmonics. The electromagnetic fields in each layer are calculated by the couple-wave approach. The boundary conditions are applied in sequence at each interface of neighboring layers to obtain the reflected and the transmitted diffracted filed amplitudes and the diffraction efficiencies [20]. The RCWA method is very efficient and powerful to calculate the diffraction efficiency and field distribution for arbitrary geometry shape of 1D and 2D periodical structures, which can be made of either pure dielectric or metallic ohmic loss material. The PFO RWG structure used in this work is pure dielectric. Since only the PFO layer has small absorption in the wavelength range of calculation, only real part of refractive index of each layer of the RWG structure was used in the calculation. In the RCWA calculation, the incident beam can set arbitrary polarization and direction [18–21]; the incident geometry can be either non-conical or conical. In the former the grating vector lies in the incident plane, while in the latter the grating vector does not lie in the incident plane. In this work only non-conical incident geometry was considered in the simulation, because the experiments were performed solely under this geometry, as shown in Fig. 1(a). In the simulations, one unit cell is defined as a single period of the PFO RWG structure with periodic boundary conditions in X direction [see Fig. 1(a)]. The number of harmonics used in simulation is 25. Other simulation parameters, shown in the caption of Fig. 3, were chosen based on refractive indices and thicknesses of the designed structure.

 

Fig. 3 (a) Calculated and measured transmission spectra of the PFO RWG for TE mode at normal incident. (b) Calculated transmittance in TE mode of the PFO RWG as a function of incident angle and wavelength (color map) and measured angular-resolved GMR modes (open circle) of the fabricated PFO RWG for TE mode. The horizontal and vertical dashed lines highlight one GMR mode at 810 nm wavelength with incident angle θ_i~32°, and another GMR mode at 550 nm, i.e. the emission peak of TPP signal of PFO polymer, with the collected angle θ_c sets at 15°. (c) Calculated electric-field intensities (E²) in the PFO RWG for the resonance modes of λ = 813 nm at θ_i = 30°. The parameters used in the calculation are Λ = 385 nm, d = 85 nm, T_TiO2 = 55 nm, T_PFO = 280 nm, and T_SU8 = 950 nm. The refractive index of glass is fixed at 1.48 for all wavelengths, but the dispersion properties of TiO₂, PFO and SU8 are used in calculations.

Download Full Size | PDF

3. Experimental results and discussions

3.1 Characterization of a PFO RWG

Figure 3(a) shows the measured and calculated transmission spectra of the PFO RWG in TE mode at normal incidence. Both the measurement and the calculation indicated that a GMR mode occurs at the downward spike at 634.2 nm. The band width of this mode is measured to be 4 nm. The measured minimum transmission efficiency is ~40%, which is higher than that in the calculation. This probably results from minor differences between the real grating structure and the ideal sinusoidal grating assumed in the calculation. Figure 3(b) shows a 2-dimensional map of the calculated transmission of the PFO RWG structure as a function of wavelength and incident angle. The GMR mode shown in Fig. 3(a) splits into two separate modes with increasing incident angle, as a result of first order diffraction from the grating: one red-shifts with the incident angle, and the other blue-shifts. Figure 3(b) displays the measured dispersion of the GMR mode of the PFO RGW structure. The measured GMR modes nearly fall right on top of the calculated low transmission tracks, indicating excellent agreement between the calculation and the measurement. Two GMR modes highlighted in Fig. 3(b) suggest that it is possible to achieve double enhancement of TPP from the PFO layer by aiming a ~810 nm wavelength (PFO two-photon absorption peak) laser beam to excite the PFO layer at 32° incident angle, and collect the PFO emission at 15° collection angle for 550nm wavelength (PFO emission peak) signal. As an example of local field enhancement, Fig. 3(c) shows the calculated electric-field intensities (E²) in the PFO RWG for the GMR mode with λ = 813 nm and θ_i = 30°. Nearly 200 fold enhancement was predicted for the “local filed” at the interface of TiO₂ grating.

3.2 Enhanced TPP via resonant two-photon excitation and enhanced extraction

Figure 4(a) shows the TPP signal collected from the PFO RWG as a function of the excitation wavelength with the incident angle of the excitation laser beam fixed at 32° and the collection angle set at 0°. The wavelength of the excitation beam was varied from 790 nm to 830 nm in 1 nm steps. The peak TPP intensity at λ = 810 nm from the PFO RWG clearly demonstrates the effect of enhancement of the excitation light due to its resonant coupling to GMR mode. To evaluate the enhancement factor, a glass slide coated with a PFO film with the same thickness is used as the control sample [5,12], and its TPP signal was measured for comparison. The TPP intensity from the PFO RWG was enhanced by a factor of 138 compared with that from the control sample. Furthermore, Fig. 4(b) shows that the excitation peak wavelength of the TPP signal at 550 nm from the PFO RWG red-shifts with the incident angle of the laser beam. In fact, the TPP excitation peak wavelengths agree with the GMR mode that red-shifts with incident angle in Fig. 3(b). Therefore, they all result from resonant coupling between the incident beam and the PFO RWG structure.

 

Fig. 4 (a) TPP intensity at 550 nm versus the two-photon excitation wavelength with an excitation beam incident angle set at 32°. (b) Normalized TPP intensity versus two-photon excitation wavelength under different incident angles: θ_i = 30°, 32°, 34°, and 36°. For these measurements, the collected angle (θ_c) was fixed at 0^ο.

Download Full Size | PDF

To further enhance the TPP signal with the PFO RWG, we next investigate the effect of enhanced extraction of the fluorescent emission. The PFO RWG sample was illuminated with 810 nm wavelength laser beam at θ_i = 32°, and the collection angle of detector was varied from 0° to 25° in one-degree steps. Figure 5 displays TPP intensity as a function of collecting angle. The TPP intensity reaches a maximum when the collected angle was set at 14°. This is the orientation where 550 nm wavelength light resonant with the PFO RWG, as indicated by one of modes that blue shifts with incident angle in Fig. 3(b). At this resonant condition the TPP intensity was enhanced nearly 220% compared with that collected at 0°. The improvement is similar to that in the OPF experiment [15], in which they observed 5-fold enhancement. Compared with the TPP intensity from the control sample, the TPP signal enhanced with both the excitation and extraction at resonance with GRM modes of the PFO RWG produces a total enhancement factor of 304. This enhancement factor is very close to that of a previous study [5], in which a nearly 330-fold enhancement was achieved. Since the RWG structure presented in this work is not yet fully optimized, further increase of the TPP enhancement ratio is possible by first reducing the thickness of the PFO layer. This can be seen in Fig. 3(c), where strong electric field of the GMR occurs to regions in the PFO film that is in close proximity of the interface between the TiO₂ grating and the PFO layer. Consequently, it can be expected that a thinner PFO layer on top of the TiO₂ grating would results in higher enhancement factor. It is reasonable to predict that a monolayer or few monolayers of PFO should also result in large enhancement of TPP, which makes the RWG structure highly feasible for two-photon bio-sensing applications.

 

Fig. 5 TPP intensity at 550 nm versus the collected angle under resonant two-photon excitation at λ = 810 nm with an excitation beam incident angle (θ_i) set at 32°. Inset shows the comparison of TPP intensity versus the two-photon excitation wavelength at two collected angles: θ_c = 0° and 14°.

Download Full Size | PDF

4. Conclusions

We demonstrated that the TPP signal can be multiplicatively enhanced up to 300-fold by aligning the GMR modes with the incident excitation light and the peak emission wavelengths of the PFO layer in the RWG structure. By setting the incident angle of the 810 nm wavelength laser beam at 32°, incident light resonates with the grating wave structure and produced strongly enhanced E-field near the TiO₂ and PFO layer interface, leading to enhanced two photon absorption of the PFO polymer and a 138-fold enhanced fluorescent signal. Furthermore, by setting the collection angle of the detection system at 14° away from the normal direction of the RWG, another 2.2-fold enhancement of the fluorescent signal at 550 nm wavelength was obtained due to the high reflection of the GMR mode. Our results indicated that such multiplicative effect is useful for further enhancement of TPP intensity of fluorescent dyes, which is highly feasible for two-photon bio-sensing applications.