Plasmonic optical fiber gratings based on few-layer Ta₂C MXenes for refractive index sensing

The TFBG was obtained by inscribing a grating on a single-mode quartz fiber using the phase-mask plate method. A high pulse excimer laser was focused onto the mask plate. Interference fringes formed a periodic RI change within the fiber core. The tilt angle of the grating was modulated by rotating the mask plate along the optical axis. Au films were deposited using vacuum thermal evaporation. A 50 nm thick Au film was deposited at a rate of approximately 7 × 10⁻⁵ Pa, 0.2 Å s⁻¹. Subsequently, the TFBG was rotated 180 degrees axially and re-deposited under the same conditions to ensure the complete coating of the Au film. Ta₂C coating was deposited using a photo-deposition method. The prepared TFBG/Au sensor was connected to the optical path, and the sensing area was immersed in a Ta₂C dispersion at a concentration of 2 mg ml⁻¹. The plasmonic local electric field led to the adsorption of Ta₂C nanosheets on the Au film, which facilitates the deposition of materials on non-planar structures. Growth spectra were recorded to analyze the optimal deposition time.

The sensor was encapsulated in a PDMS-based sensing cell (width: 2 μm, height: 2 μm, length: 3.8 cm) to facilitate interaction with the analyte. A microfluidic peristaltic pump (QHZS-001B) was utilized to facilitate the injection of the analyte and control of the flow rate. A hybrid light source (ASE-CL-30-M, SLD-1470-10-B) was adopted to ensure a wide spectral band covering the detection band of the sensor. The spectra were analyzed by a high-resolution spectrometer (Yokogawa AQ 63700). Different concentrations of ethanol were taken as the analytes. Their RIs were modulated on demand using an Abbe refractometer (DR-A1).

The microscopic morphology of the sensor surface Au film and Ta₂C were conducted using scanning electron microscopy (SEM, Zeiss Gemini Ultra-55) and TEM (JEM-2100F). The thickness of the Ta₂C nanosheets was determined by atomic force microscope (AFM, Horiba SmartSPM). The lattice features of Ta₂C were analyzed by HRTEM and SAED. XPS (Thermo Fisher Scientific 250Xi), Raman spectroscopy (Horiba HR Evolution), and energy dispersive spectroscopy were employed to determine the composition and valence of the plasmonic structure in the sensing region. The Raman spectrum was obtained using a 532 nm exciting laser, with 1200 grooves per 1 mm, ×50 objective lens, 1 μm laser spot, and 8 s of acquisition time.

The greatest advantage of TFBG spectrum is the dense comb-like peaks. The SPR mode is supported under the Au layer attachment. Typical TFBG-SPR transmission spectra are shown in figure 1(a). The core mode propagating within the fiber core results in a signal peak located at the spectral Bragg wavelength. Since it responds more to the changes in temperature and stress of the fiber and less to the external environment, it provides a benchmark for avoiding errors induced by interfering factors [36]. The reflection of the light beam by the grating caused part of the light to be coupled into the cladding to form a cladding mode; the cladding modes with different effective indices were manifested as peaks of different wavelengths in the comb spectrum. Owing to the coating of the Au film, the specific cladding mode can be coupled to the SP at the Au/cladding interface, expressed as the SPR peak at the wavelength corresponding to such an effective index [37]. Figures 1(b), (c) shows the cross-sectional electric field distribution and field profiles for bare TFBG and Au-coated TFBG. The SPR supported by the Au-coated TFBG generates a strong electric field, which causes a significant increase in the energy density outside the fiber, especially at the upper and bottom ends of the fiber. The SPR wavelength is intimately related to the RI of the surroundings, demonstrating a red shift of the peak with the increase in the effective index [38]. In addition, the wavelength and amplitude of the comb peak shift with RI change, as shown in figure 1(d). The amplitude is extremely sensitive to small RI changes and the wavelength can be effective over a larger RI range. The wavelength shift of the signal peak, in this case, can be described as:

$$\begin{eqnarray}&&{\rm{\Delta }}{\lambda }_{clad}^{j}=\left(\displaystyle \frac{\wedge }{cos\theta }\displaystyle \frac{\partial \left({n}_{core}^{j}+{n}_{clad}^{j}\right)}{\partial {n}_{SRI}}+\displaystyle \frac{{n}_{core}^{j}+{n}_{clad}^{j}}{cos\theta }\displaystyle \frac{\partial \wedge }{\partial {n}_{SRI}}\right){\rm{\Delta }}{n}_{SRI}\end{eqnarray} \tag{ 1 }$$

where ${\rm{\Delta }}{\lambda }_{clad}^{j}$ and ${\rm{\Delta }}{n}_{SRI}$ denote the wavelength shift of the j-order cladding mode and RI change, respectively; $\wedge$ and $\theta$ indicate the grating period and grating tilt angle, respectively; ${n}_{core}^{j}$ and ${n}_{clad}^{j}$ represent the effective indexes of the core mode and cladding modes, respectively. Since $\wedge$ and ${n}_{core}^{j}$ are insensitive to changes in RI, there are: [39]

$$\begin{eqnarray}&&\partial {n}_{core}^{j}/\partial {n}_{SRI}=0\end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray}&&\partial \wedge /\partial {n}_{SRI}=0\end{eqnarray} \tag{ 3 }$$

An approximation of equation (1) yields the following relationship: [40]

$$\begin{eqnarray}&&{\rm{\Delta }}{\lambda }_{clad}^{j}=\displaystyle \frac{\wedge }{cos\theta }\displaystyle \frac{\partial {n}_{clad}^{j}}{\partial {n}_{SRI}}{\rm{\Delta }}{n}_{SRI}\end{eqnarray} \tag{ 4 }$$

This allows the detection of RI changes by analyzing the shift of the resonance wavelength. Figure 1(e) shows the linear variation of the RI-SPR dip, demonstrating the potential for quantitative detection. Notably, the cladding modes with different effective indices have different sensitivities to the RI response. Therefore, the wavelength sensitivity and figures of merit (FOM) of signal peaks 1–5 were analyzed to determine the optimal signal peak. Figure 1(f) presents the performance of each signal peak measured regarding wavelength sensitivity and FOM. The second peak has the best sensing performance in these metrics. Given the above results, peak 2 was selected to measure the sensor performance in the subsequent experiments.

Zoom In Zoom Out Reset image size

Ta₂C thin film was grown on the surface of the Au film to form an Au/Ta₂C hybrid plasmonic structure and hence further improve the plasmonic photoelectric response of the sensor. A photo-induced deposition method was employed to obtain high-quality few-layer Ta₂C film. The optimal enhancement conditions were explored by varying the growth time to modulate the film thickness. Figure 2(a) exhibits the surface and cross-sectional morphology of bare Au and Au/Ta₂C structures with growth times of 60–180 min, respectively. Concerning the pure Au sensor structure, the surface was a uniform and flat Au film with a thickness of 50 nm. With the growth time increasing to 60 min, the Au film surface gradually became mottled, and the cross-section presented a sporadic distribution of Ta₂C nanosheets. The markings in figure 2(a₂) confirmed the growth of Ta₂C. As the growth time increased to 120 min, the shadow of Ta₂C appeared in a large area, and clear, continuous Ta₂C films appeared in the cross-sectional view in figure 2(a₃). Besides, a white agglomeration appeared on the sensor surface when the growth time reached 180 min. The cross-sectional view in figure 2(a₄) demonstrates the stacking and agglomeration phenomenon caused by Ta₂C over-deposition. This Ta₂C agglomeration was up to several hundred nanometers thick and formed a macroscopic granular structure. The sensing spectra of each structure were plotted to measure the specific sensing performance. The increase in Ta₂C thickness was visualized in the detection spectrum as a redshift in the resonance wavelength. However, the increase in sensitivity was not monotonically with an increase in the thickness of Ta₂C. Figure 2(b) compiles the variation of wavelength sensitivity and FOM with growth time. The increase in sensitivity from 0 to 120 min growth time corresponded to the favorable growth of Ta₂C. The sensitivity and FOM reached 229.5 nm/RIU and 2228.15 at 120 min, which was a significant improvement relative to the uncoated sensor. The sharp decrease in sensitivity at 180 min was induced by the over-thick Ta₂C film that hindered the interaction of the sensor with the external environment. The additional energy loss introduced by the disordered stacking of Ta₂C suppressed the SPR, leading to the insignificant SPR signal spectrum in the inset of figure 2(a₄). Similarly, the resonance peak was finely red-shifted after 120 min due to the over-thick growth of the Ta₂C film. Spectral changes during growth (figure 2(c)) were also recorded to measure the growth of Ta₂C. The resonance wavelengths of the key signal peaks exhibited a redshift with time. The amplitude of the resonance peaks also gradually increased with time. After about 120 min of growth, the change in the key wavelength shift became weak and only changed in amplitude, which was a disturbance caused by the thick coating [41]. The shift of the SPR wavelength was minor throughout the growth process. This phenomenon can be attributed to the decrease in RI derived from the gradual deposition of Ta₂C dispersion. The spectrum of the TFBG/Au/Ta₂C sensor with the ideal morphology and the optimal sensing performance is plotted in figure 2(d) and has a 1.2 nm shift in resonance wavelength relative to the TFBG/Au sensor. This enhancement was confirmed by subsequent characterization and analysis.

Zoom In Zoom Out Reset image size

The Au/Ta₂C structure was characterized in various aspects for subsequent analysis. The transmission electron microscope (TEM) image in figure 3(a) demonstrates the favorable deposition of Ta₂C on Au film. High-resolution transmission electron microscopy (HRTEM) images (figure 3(b)) presented the [1010] crystalline surfaces of the grown Ta₂C and had a good crystalline state where the planar spacing of Ta-Ta and Ta-C were 2.8 Å and 2.5 Å, respectively [42, 43]. The inset displays the selected area electron diffraction (SAED) pattern in the Ta₂C [0001] axial direction. The developed [$10\bar{1}0$], [$01\bar{1}0$], and [$11\bar{2}0$] planes suggested that the grown layer conformed to the MXene phase [44]. AFM was employed to analyze the thickness of the Ta₂C nanosheets. The height profile in figure 3(c) indicates that the thickness of the nanosheets is approximately 3.3 nm. Subsequently, x-ray photoelectron spectroscopy (XPS) was adopted to qualitatively analyze the elemental composition and chemical state of the Au/Ta₂C structure. A high-resolution spectrum of the Ta 4 f region is depicted in figure 3(d), which reveals a dominant oxidation state characteristic peak of Ta⁵⁺ at 26.1 eV, as well as a relatively pronounced oxidation state characteristic peak of Ta⁴⁺ present at 23.4 eV. In addition to the 4f eigenpeak present at 22.4 eV, two eigenpeaks at 24.6 eV and 28.1 eV arising from the TaC_x 4f_5/2 and TaC_xO_y 4f_5/2orbitals were also present [45]. These peaks verified that a unique MXene phase was obtained. Then, energy dispersive spectrometer (EDS) and Raman spectroscopy were employed to characterize the coverage of the Ta₂C. EDS was utilized to demonstrate favorable deposition of Au and ideal growth of Ta₂C in terms of elemental content. The energy dispersive spectrum of figure 3(e) determines the simultaneous presence of Au and Ta elements. Besides, the Raman spectrum in figure 3(f) exhibits representative characteristic peaks at around 237, 651,853,1342, and 1577 cm⁻¹. The peak located at 237 cm⁻¹ can be attributed to the in-plane and out-of-plane vibrations of Ta atoms in the outer layer as well as C and surface groups. The other peaks are attributed to C vibrations. The uniform presence of Ta₂C is demonstrated by the Raman mapping image of the 1577 cm⁻¹ peak [45, 46]. These ideal structures guarantee the enhancement of the sensor performance, as analyzed below.

Zoom In Zoom Out Reset image size

The enhancement mechanism of Ta₂C was analyzed by the energy band theory. Figure 4(a) illustrates the charge exchange and energy levels in the sensor. The introduction of Ta₂C leads to the blocking of the interband transition, manifesting as a decrease in the energy required for electron transfer at the Fermi energy level. Thus, the free electrons were more easily transferred from the Fermi energy level of Au to the hole in the affinity energy level of the Ta₂C when a specific wavelength incidence excited the surface plasmon of Au. Subsequently, the relaxation instant electrons returned to the holes of Au. This process was accompanied by excited radiation, which enlarged the real part of the effective RI and amplified the polarizability response of the system [47]. The energy band map of the 5-layer stacked Ta₂C bulk phase was calculated by Vienna Ab-initio Simulation Package (VASP). Figure 4(b) exhibits the high symmetry points in the Brillouin zone of the Ta₂C film. The multiple intersections at the Fermi energy level of the plot under this path (Γ-M-K-Γ-A-L-H-A) corroborated the metallicity of Ta₂C. Conventional 2D materials represented by graphene are semi-metallic in nature and have a low carrier concentration. Relative to other 2D materials, Ta₂C implied stronger carrier migration efficiency, wider plasmonic frequency, and high electric field confinement properties, and enhanced sensing capability [31, 48]. Afterward, the magnitude of the electric field enhancement of the structure was simulated in response to the SPR intensity. Figures 4(c), (d) demonstrates the electric field intensity distribution of the Au-coated and Au/Ta₂C-coated TFBG. The symmetrical stripe-like pattern produced in the fiber axial direction indicated the propagation of the SP wave. The Au/Ta₂C sensor possessed an enhanced electric field intensity and distribution range, verifying the SPR enhancement effect [49]. Additionally, the plasmonic absorption resonance spectra (figures 4(e), (f)) of Au and Au/Ta₂C systems were simulated to evaluate the RI sensing capability. The Au sensor exhibited a wide absorption peak at 1544 nm. The wavelength was redshifted by 0.84 nm as the RI changed from 1.3360 to 1.3365. The sensitivity was calculated to be 168 nm/RIU. The absorption spectrum of the Au/Ta₂C sensor was shifted by about 0.6 nm relative to the Au sensor. However, Ta₂C contributed to an enhancement of the resonance peak intensity, as well as a narrow signal peak, which corresponded to an enhancement of the FOM. Furthermore, the resonance peak redshift and sensitivity for the same RI change interval were 0.95 nm and 190 nm/RIU, respectively. Theoretically, Ta₂C provides an improvement in RI sensing capability.

Zoom In Zoom Out Reset image size

Melamine, a carcinogenic substance, was once illegally added to dairy products because of its extremely high nitrogen content. Besides, analytes were formulated with different melamine concentrations in milk. Thus, melamine was selected as the molecule for the comprehensive analysis of the RI sensing performance. The optical path shown in figure 5(a) was used for the actual detection. The broad spectrum generated by the broadband light source (BBS) was transformed into polarized light by a polarizer and polarization controller (PC). The sensing region of the fiber was subsequently encapsulated by a sensing cell for sensing in liquid environments. The signal was finally collected by an optical spectrum analyzer (OSA). The operation process of the sensing area is illustrated in figure 5(b). The Raman spectra of the melamine in milk and water are provided in figure 5(c). The most significant peaks involve the in-plane deformation of the triazine ring. It shifted from 674 cm⁻¹ to 700 cm⁻¹ due to the change in the solution environment, which is consistent with previous studies [50]. Figure 5(d) exhibits the spectra of melamine with concentrations from 10⁻² to 10⁻¹¹ M. The decrease in melamine concentration led to a change in RI. Hence, the wavelength and amplitude of the resonance peak decreased in the spectra. However, the change in RI due to the decrease in the concentration of melamine to 10⁻⁹ M was minimal. This phenomenon made the spectral changes indistinguishable, indicating that the limit resolution of the sensor was approached. For such a small refractive index variation (less than 10⁻⁵ RIU), interrogating the amplitude is an effective sensing approach. Figure 5(e) specifies the variation with concentration on the amplitude. There is a favorable linear response between the amplitude and the concentration of the analyte. As suggested in the standard curve, the sensor possesses excellent linearity over the 10⁻³ to 10⁻⁸ M concentration range with an R² value of 0.990. The limit of detection (LOD) can be calculated by the equation LOD = 3 SD/S⁻¹, where SD represents the standard deviation of the response and S refers to the slope of the calibration curve. The SD of the response at 10⁻⁸ M is 0.0281, thus the LOD was calculated to be 7.9 × 10⁻⁹ M, which was in accordance with the internationally established of 1.5 × 10⁻⁴ M.

Zoom In Zoom Out Reset image size