Tuning Photoluminescence of Biological Light Emitters via Silk Protein Based Resonators

Abstract

Adding tunability to biological light emitters offers an unprecedented technique in biological sensing and imaging. Here, we report a tunable, lithographic-free, planar, and ultrathin metal-insulator-metal (MIM) resonator capable of tuning the optical properties solely by a silk/sodium fluorescein hydrogel layer, a biocompatible light emitter. In water, the volume of the resonator was expanded by swelling, and then the resonant mode could be shifted. Simulations predicted the red-shifted resonance peak in transmission when the MIM was swollen in water. The red-shift could be attributed to the increase in the thickness of the silk hydrogel layer due to the absorbed water. The shift of the resonance could affect the fluorescence of the dye in the silk hydrogel layer.

I. INTRODUCTION

Photoluminescence (PL) tuning is an important subject due to its significant applications in lasers, telecommunications, biomarkers, and medical imaging [1, 2]. PL spectra have been used to study the characteristic features, even impurities of fluorescent substances via several light-emitting metal-structures [3]. Tuning of the photoluminescence has been achieved through the alteration of aggregated structure within various devices such as quantum dot light-emitting devices [3], metamaterials with Fano resonance [4] and piezoelectric single-crystals [5]. These devices no doubt, are operative for fluorescence tuning, but the involved experimental progressions are complex and the non-economical fabrication techniques limit the applications.

To overcome all these problems, ultra-thin multilayered planar resonators are a better choice especially due to the delicate control over the spectral properties and the manipulation of light-matter interactions. The flexibility of the spectral engineering and design of these planar and multilayered optical structures give the advantages of cost-effective fabrication, high sensitivity, selectivity and availability [6-10]. Moreover, the exciting optical functions such as optical resonances [11], field localization [12], refractive index (RI) engineering [13], phase and amplitude control [14], etc., are achieved simply by modifying the size, shape, and dielectric material of these planar devices[15-18]. For instance, highly efficient spectral properties of the planar metal-insulator-metal (MIM) resonators, based on the Fabry-Perot etalon principle, have been used to tune the PL spectrum.

Here, we report a planar ultrathin tunable transmission color filter based on metal-insulator-metal (MIM) resonators capable of tuning the optical properties solely by the film thickness and RI of the insulating material. Silk fibroin,the natural protein extracted from the Bombyx Mori, has emerged as an excellent dielectric material in optics and light manipulating fields due to its optical transparency and unique mechanical behavior [19-21]. It has been used in various devices such as photonic crystals [22, 23], waveguides [24, 25], and single-mode lasers [26, 27]. The hydrogel form of silk is capable of inducing drastic changes in the optical resonances due to its ability to swell and shrink [28]. These properties of silk in MIM resonators can tune the spectral outcome. The intermediate silk hydrogel layer serving as a dielectric between two metal layers is mixed with sodium fluorescein. The optical pumping can excite the electrons of sodium fluorescein in the film sandwiched between two metal films and upon de-excitation,it emits photons that exhibit a PL spectrum. The objective of this study was to fine-tune the PL spectrum of sodium fluorescein by altering the thickness and refractive index of silk in an MIM resonator. This is achieved by simply water absorbance.

II. METHODS

Preparation of silk aqueous solution: The cocoons of bombyx mori caterpillars were boiled in a 0.02 M NaCl solution. The fibroin was rinsed with distilled water and air dried for 24 h. The dried fibroin was dissolved in a9.3 M LiBr solution for 4 h at 60°C, yielding a 20 wt%aqueous solution. This solution was dialyzed to 6% in the cellulose tubular membrane for 2 days at room temperature followed by centrifugation and syringe filtration with a0.45 µm pore size.

Fabrication process: The bottom silver layer 30 nm thick was deposited on the square quartz substrates of 2 × 2 cm2using an electron beam evaporating process. An 8 wt% of sodium fluorescein salt solution was mixed in 6 wt% silk aqueous solution. This solution was coated on the silver layer by using a spin-coater up to 130 nm thickness. The thickness of the blended silk spacer could be controlled by the concentration of silk and by the revolution of the spin coater which was set to 5.77% and 3900 rpm. To convert silk/sodium fluorescein solution into the hydrogel, the sample was dipped into methanol for 1 min and air dried. Then the top layer of silver of 30 nm was deposited on the silk/sodium fluorescein hydrogel by the same electron beam evaporation method.

Simulation: To perform the numerical simulations, we used a commercially available finite-difference time domain(FDTD) software package (FDTD, Lumerical Solutions).Simulations were carried out in the 2D layout. The 300 nmunit cell was selected along the x-axis and the boundary conditions were imposed on the lateral directions. The plane wave source was placed inside the MIM device along the +y direction and for comparison purpose it was moved outside of the device. The wavelength range for this plane source was 300~1000 nm with 2000 frequency points. The transmission was recorded by means of the power monitor below the silver layer. The 2D field profile monitors in x-y plane showed the electric and magnetic field distribution across the cross-section of the MIM. The complex refractive index of Ag and silk and quartz was taken from the data of Palik [29].

Optical measurement: For transmission measurements, optic fibers with a 400 µm core diameter were used, one for directing and the other one for collecting the light from the sample. The white light source was placed in the vertical direction, while the transmitted signal was displayed by using an optical spectrum analyzer via an optic fiber. The MIM samples were pumped by the continuous wave from the Nd: YAG laser. The diameter 0.25 µm incident beam was focused on the sample with the help of lenses. The resulting signals were collected by optic fiber and directed to an optical spectrum analyzer.

III. RESULTS

The MIM device was fabricated with e-beam silver deposition and followed by spin coating of the silk layering. FIG. 1(a). Silk protein mixed with sodium fluorescein was sandwiched between top and bottom ultrathin Ag layers(30 nm each) and this insulating layer was treated with methanol to induce β-sheet crystallization through the formation of hydrogen bonds between silk molecules, which transforms the soluble silk solution into the hydrogel [9,10, 30]. This MIM resonator mimics the Fabry-Perot etalon principle that depends upon the variations of thickness and refractive index of the insulating layer resulting in different resonance peaks Fig. 1(b).

FIG. 1. (a) The fabrication process of the MIM resonator. (b) The schematic working principle of the MIM device. The measured spectra of silk mixed with sodium fluorescein MIM in the dry and wet state are shown. Note that the silk in the expanded state can tune the transmission spectra.

FDTD simulations were carried out by assuming a light source (in the range of 400~1000 nm wavelength) first positioned inside the dielectric (silk/sodium fluorescein) of the MIM and then moved outside of the MIM resonator for the comparison. These simulations were studied in the presence of air and water medium sequentially (Fig. 2(a)). The resonance peak is red-shifted due to the swelling of silk in the water medium irrespective of the light source position. In the air medium, the wavelength is shorter and is possibly due to the dry state of silk hydrogel as compared to for the water medium. Again, the position of the light source makes no difference, while the thickness of the dielectric material is the determining factor in the output transmission spectra. The FDTD simulation also predicted the change in the thickness of the silk/sodium fluorescein layer from 115 nm in the air medium to 190nm as a result of water absorbance. To validate these simulation results, the experimental setup with the light source placed outside was established to record the transmission spectra of the silk/sodium fluorescein layer in the MIM device. In the air medium, the dry state of the silk/sodium fluorescein layer transmits a resonance peak of wavelength ~530 nm while in the water medium, thered-shifted wavelength ~695 nm was observed, as shown in Fig. 2(b). These experimental results are in strong agreement with those of simulation predictions when the source is placed outside. However, for the placement of the light source inside the MIM cavity, the experimental setup is not viable due to the nano-dimensions of the MIM resonator. Therefore, to further investigate the effectof inside light source positioning, more FDTD simulations were performed varying the transmitted electric field (TE)and the magnetic field (TM) intensities.

FIG. 2. (a) FDTD simulation results for transmission with the source inside and outside. (b) Experiment results for transmission peaks in the air and wet state with a source outside of MIM.

The stimulated TE and TM intensities transmitted through air and water media in relation to the light source position inside and outside of the three-layer MIM cavity are shown in Fig. 3. In both media, the electric field is highly confined at the dielectric section between two metallic films, where a standing wave is formed due to the constructive interference of incoming and reflected waves. In air and water media, transmitted wavelengths of ~526 nm and 695nm were recorded, respectively, regardless of the light source position. In short, the effect of the medium is the red-shift, which is a result of the increase in the thickness of the MIM hydrogel cavity.

FIG. 3. Electric (TE) and magnetic field (TM) confinement at corresponding transmission wavelength when the source is outside in air (a,b), and in water (c,d). When the source is outside the cavity, in air (e,f) and in water medium (g,h).

To study the photoluminescence of the sodium fluorescein, optical pumping was done with the Nd: YAG laser of λ = 355 nm as shown in Fig. 4(a). The incidence angle of the laser beam on the resonator was fixed at 0°, and the diameter of the laser spot on the film was 0.25 µm.The sodium fluorescein absorbs the incoming radiation from the pumping source which excites the electrons from ground state to excited state then the excited electron emits the photon and jumps back to the initial state. The emitted photons exhibit the PL spectrum. The PL spectrum in the expanded state of silk is blue-shifted by ~22 nm as the resonance peaks in air and water are at ~537 nm and ~515 nm respectively (Fig. 4(b)). The fluorescent molecule emits light in random directions. It bounces back and forth between two metal layers and exhibits constructive interference at some specified angle. This angle is a crucial factor in tuning the PL spectrum and results in a blue-shift in the water medium when the thickness of the MIMcavity has increased.

FIG. 4. (a) Vertical laser setup for optical pumping. (b) Display of PL spectrum of Sodium Fluorescein in air and water media.

IV. DISCUSSION

Since the MIM cavity follows Fabry-Perot-like behavior, only the small proportion of the incident light wavelength is trapped and transmitted which is selected via the resonance wavelength of the MIM cavity. This determining resonance wavelength of the MIM cavity follows the following equation.

\(2\left(\frac{2 \pi}{\lambda}\right) n d \cos \theta+\phi_{a}+\phi_{b}=2 \pi m\)       (1)

where λ, n, θ and d correspond to the resonance wavelength, the refractive index, the incidence angle, and the thickness of the MIM dielectric, respectively. The terms φaand φb show the phase shift due to reflection from the bottom and the top metal layers, respectively. The order of the MIM cavity mode is an integer m. For the light to be trapped inside the MIM cavity, the phase condition (2 πm)in Eq. (1) must be satisfied, which means to enable the constructive interference in each full round trip, the lefthand side of the equation should be an integer multiple of2π. Finally, constructive interference makes the enhanced transmission of light through the cavity possible [31, 32].

\(\left(\frac{4 \pi}{\lambda}\right) n d \cos \theta+2 \phi=2 \pi m\)       (2)

From Eq. (2), the wavelength of the peak transmission is

\(\lambda=\frac{2 \pi n d \cos \theta}{\pi m-\phi}\)       (3)

From Eq. (3) the position of the peak transmission depends on the angle of incidence such that maximum λoccurs when θ = 0 for a specified thickness and refractive index of the medium. Increasing the medium thickness fora particular value of θ, the peak transmission yields ared-shift. The peak transmission shows blue shift with increasing the incident angle θ for a specified thickness. For tuning the resonance behaviors, the volume change of the dielectric layers is crucial. The tuning of the PL spectrum depends upon the incident angle inside the cavity, RI, and the width of the cavity. When the MIMresonator is immersed in water, the silk absorbs the water molecules thus increases the thickness of the dielectric which changes the position of the fluorescein molecules.As the fluorescein molecule emits light in random directions and the change in position of molecules induces the electric field between the silver layers at some specified angle, and hence shows the blue-shift in resonance wavelength of the PL spectrum. For some specified θ, increase in thickness d ~80 nm and decrease in refractive index of silk from nsilk = 1.54 to nsilk-swelled = 1.43, Eq. (3) yields the shorter wavelength. The tuning of the PL spectra also depends upon the incident angle along with the thickness and RI of the spacer.

V. CONCLUSION

Our experimental results and theoretical explanation indicate that the dimension of the MIM cavity and the incident angle of light enable the tuning effect of the PL spectrum. In this study, post casting of the silk mixed with sodium fluorescein serving as an intermediate layer for film formation on the fabricated MIM resonator yielded a tunable MIM resonator. The FDTD simulations for the light source placed inside and outside in the dielectric silk/sodium fluorescein (cavity) of the MIM resonator, for both air and water media, were performed. Irrespective of the light source position, the simulation results predicted thered-shifted resonance peak in a water medium, which was a consequence of the increase in thickness of the silk layer. For transmission spectra, experimental results match well with those of simulation. The PL spectrum was tuned in the water medium due to the increase in the thickness of silk hydrogel and the angular dependence of the emitted light from the sodium fluorescein between the insulating layers.