Open Access
Add to BibSonomy
Abstract
Si nanopillar (NP) arrays are investigated as refractive index sensors in the visible/NIR wavelength range, suitable for Si photodetector responsivity. The NP arrays are fabricated by nanoimprint lithography and dry etching, and coated with thin dielectric layers. The reflectivity peaks obtained by finite-difference time-domain (FDTD) simulations show a linear shift with coating layer thickness. At 730 nm wavelength, sensitivities of ~0.3 and ~0.9 nm/nm of SiO2 and Si3N4, respectively, are obtained; and the optical thicknesses of the deposited surface coatings are determined by comparing the experimental and simulated data. The results show that NP arrays can be used for sensing surface bio-layers. The proposed method could be useful to determine the optical thickness of surface coatings, conformal and non-conformal, in NP-based optical devices.
© 2017 Optical Society of America
1. Introduction
In the field of bio and gas sensing [1–6], there is a high demand for sensing devices that are cheap and have high sensitivity, selectivity and reliability. Various sensing approaches have been investigated such as optical [1], electrochemical [7,8], piezoelectric (or mechanical) [9,10], and thermal (or calorimetric) [11,12]. Optical methods provide an easy, reliable, selective and highly sensitive sensing [3]. The detection principles in optical methods are often based on determining fluorescence, luminescence or spectral shifts (color change).
One type of optical sensor relies on sensing the change in refractive indices (RI). A number of approaches for this type of sensing have been reported including photonic crystal cavities [13,14], slot-waveguides [15,16], ring resonators [17,18], (Mach Zehnder) interferometry [19,20], Surface Plasmon Resonances (SPR) [21,22], and (semiconductor) nanowaveguide arrays [23,24]. In general, a peak shift change is measured related to a (slight) RI change in the vicinity of the sensing device and its sensitivity is pre-dominantly given in Refractive Index Sensitivity units (RIS; nm/RIU), which is the relation between the peak shift (Δλ) and the change in Refractive Index Unit (RIU). The detection limit (DL) is determined by the sensitivity of the sensing device and the detection resolution (R) of the detector; this value is given in units of RIU−1. The aim is to develop RI sensing devices that are highly sensitive for small RI changes and/or show selectivity for the desired sensing material. An approach to obtain a highly selective sensing device is by creating specific surface conditions, e.g., by surface treatment by which only specific molecules are able to attach to the surface [2]. Already high RI sensing sensitivities have been reported for which RIS values have even been shown in the order of RIS≈7000 nm/RIU (SPR; prism coupled) [25] and DL values in the order of DL≈10−5-10−7 RIU−1 (SPR) [1]. Most of these reported high sensitivities are obtained for either ‘complex’ devices, a limited RIU range and/or for only optical sensing in the NIR region. Though high sensitivities are shown, these devices often do not show the highest practicality for applications with regard to fabrication, large-scale implementation, material type and low-cost.
Refractive index sensing methods are also relevant for optical/optoelectronic devices based on nanostructure geometries such as nanopillars, wires and cones. Most often such structures have coating layers, e.g. either dielectrics or transparent conductive oxides, whose thicknesses and refractive indices affect the optical properties of the structures. Typically such coatings are provided by plasma based methods, sputtering or atomic layer deposition (ALD). Depending on the structure geometry and the deposition method the coatings may or may not be conformal. Thus, determination of effective optical thickness of the coating becomes a very useful parameter in device design.
Here, a sensing method is proposed based on Si nanopillar (pillars acting as a vertical waveguide) arrays which provide a high surface to volume ratio as well as possibilities to engineer the guided modes to be more surface sensitive. Specifically, we examine its suitability in the visible/NIR wavelength range, suitable for Si photodetector responsivity. Surface functionalization can be applied for obtaining a high selectivity for, e.g. bio sensing applications [26,27]. This type of structure/approach has previously already been shown to be suitable for highly selective and sensitive bio sensing applications [28]. By tuning the material, geometrical and the periodical properties of the pillar arrays, a characteristic optical response spectrum can be developed consisting of desired reflection/transmittance resonances which are used for the optical sensing. Different types of material, e.g., ZnO [29,30] and SU-8 [31,32] could be used as sensing materials, however Si is attractive from the point of view of integration and has already been widely used for making optical sensors [2,28,33]. The associated advantages are low cost, mature fabrication/processing technology, non-toxicity, and suitable electrical and optical properties.
In this work, Si nanopillar array structures with a hexagonal period of ~530 nm, a height of ~1350 nm and a diameter (top-bottom) of ~200-350 nm, are used for sensing applications in the visible/NIR spectrum. Though the highest sensitivities are pre-dominantly reported for the NIR region, here a sensing option in the visible/NIR region is assessed as a more practical method; related to the light source and detector type. Pillar array structures were fabricated by nanoimprint lithography (NIL) which has the advantages of scalability and low-cost. However, alternative low-cost approaches such as colloidal lithography (CL) could also be attractive. For the NIL fabricated structures, electromagnetic (EM) modeling and simulations and optical characterization of the reflection spectrum have been performed in order to investigate its sensing capabilities. The spectral shifts of the characteristic reflection peaks have been related to RI changes of either the medium between the pillar structures or to local RI changes due to a (thin) RI material layer coverage on the pillar structures. For the change of the RI of the medium between the pillars, the theoretical/simulated RIS is determined. For the RI layer sensing of the (thin) RI layer on the Si pillar structures, the reflection peak shift is related to the RI and its layer thickness. This method provides a way for either determining the RI when the (average) layer thickness and peak shift is known or to determine the (average) layer thickness when the RI and peak shift is known.
In addition, this (semiconductor) nanopillar array sensing method provides a flexible method for which the characteristic reflection spectrum can be tuned by optimizing the material, period and geometry of the pillar arrays for the specific application purpose with regard to the: abundancy of the material, specific spectral region for sensing, ease of fabrication, cost reduction and large-scale implementation. The optical RI sensing approach presented here provides a cheap and straight forward method for determining (localized) RI changes in the vicinity of the (Si) nanopillar/wire array structures.
2. Fabrication
The Si NP arrays were fabricated using nanoimprint lithography (NIL) and dry etching. The NIL pattern consisted of a hexagonal array of circular resist patterns. For better etch selectivity, the Si (100) wafers were covered with a 100 nm SiO2 layer deposited by plasma enhanced chemical vapor deposition (PECVD). The deposition parameters were: 2%SiH4/N2 at 710 sccm, N2O at 425 sccm, temperature of 300 °C, RF power of 20 W, pressure 800 mTorr and a planar (calibrated) deposition rate of ~1.1 nm/s. After NIL, the residual resist removal was done by a mild O2 plasma treatment (O2 flow of 50 sccm, RF power of 100 W, pressure of 50 mTorr and an etch time of 30 s). The NIL pattern was transferred into the (mask) SiO2 layer by reactive ion etching (RIE; CHF3 flow of 25 sccm, RF power of 100 W, pressure of 15 mTorr and sufficient etch time). The sample was soaked in acetone for 1 hour to remove the resist. The patterned SiO2 is used as the etch-mask to fabricate Si NP arrays by an ICP-RIE etching using a pseudo-bosch process. The process parameters were: C4F8 flow (passivation gas) of 80 sccm, SF6 flow (etching gas) of 35 sccm, RF platen power of 10 W and ICP power of 575 W. After Si etching, the residual SiO2 mask was removed using HF. The obtained NIL Si NP array structure (NIL1) has a hexagonal period of ~530 nm, a pillar height of ~1350 nm, and a top and bottom diameters of ~200 and ~350 nm, respectively. A representative SEM image of the fabricated NP array is shown in Fig. 1(a).
Fig. 1 (a) Representative SEM image (tilt 25°) of the fabricated Si NP 530 nm period hexagonal array. Si nanopillar is truncated conical in shape with a pillar height of ~1350 nm and a diameter (top-bottom) of ~200-350 nm. (b) Representative SEM image (tilt 25°) shows a NP array with a SiO2 over-layer of ~20 nm average thickness.
To characterize surface layer (refractive index) sensing, two different dielectric layers (SiO2, n≈1.5 and Si3N4, n≈2) with different refractive indices (RI) were deposited by PECVD on the Si NP arrays. As optical sensing principle, the peak shift(s) of the resonant peaks of the characteristic reflection spectrum of the NP array is used to relate this shift to the (average) layer thickness (δ) and the effective RI of the layer (nlayer). By measuring the (average) layer thickness of the RI layer and its related peak shift, the effective RI of the layer can be determined; and vice versa.
Different average layer thicknesses of SiO2 and Si3N4 varying between ~0-30 nm have been investigated. The PECVD settings for the SiO2 layer deposition were: 2%SiH4/N2 flow of 710 sccm, N2O flow of 425 sccm, temperature of 300 °C, RF power of 20 W, pressure 800 mTorr and a planar (calibrated) deposition rate of ~1.1 nm/s; and for Si3N4 were: 2%SiH4/N2flow of 800 sccm, NH3 flow of 16 sccm, a temperature of 300 °C, RF power of 24 W, pressure of 650 mTorr and a planar (calibrated) deposition rate of ~0.3 nm/s. Before each layer deposition, the NP array sample was cleaned with standard organic solvents by sonication, rinsed with DI water and blow dried by N2. In the final cleaning step O2 plasma surface treatment was used for 10 min with an O2 flow of 500 sccm and RF power of 1000 W, respectively. After optical characterization, the respective dielectric layers were stripped using 50% HF.
SiO2 and Si3N4 PECVD layers on the NPs are not conformal and non-uniform coverage is typical at relatively high process pressures, with thicker coatings at the upper portions of the NPs. The measured SiO2 and Si3N4 layers on planar Si samples under same deposition conditions, as expected showed higher thickness and follow the calibrated deposition rates. Due to the significantly lower process pressures and the deposition rates, Si3N4 films can be expected to be more conformal compared to SiO2. Nevertheless, the effective “optical” thickness can be determined by comparing the simulations and experimental reflectivity peak shifts. However, roughness at the top of the pillars and the relatively higher accumulation of the deposited material at the pillar top can influence the measured reflection spectrum. Scanning electron microscopy (SEM) was used to estimate the (average) layer thicknesses on the NPs. A representative SEM image of the deposited layer is shown in Fig. 1(b). Before the deposition of a new layer, the (average) diameter of ten separate Si pillars has been determined at the top, middle and bottom of the pillars using SEM imaging. These values have been averaged separately and were used as a reference for the next layer deposition. After each new layer deposition the average diameter change has been determined, in a similar way, and from this the overall average layer thickness was derived and was used in the peak shift analysis. Although the procedure can have errors, especially for thin layers, it provides a rough estimate.
3. Finite–difference time-domain (FDTD) simulations
FDTD simulations using a Lumerical tool have been performed to investigate the sensitivity for (localized) RI sensing. In addition the volume sensing capability has been assessed. The characteristic reflectance spectrum of the NIL1 is the key feature for this. Ideally, the reflectance spectrum should show sharp single peak(s), which can easily be resolved for small peak shifts, is of appreciable intensity and is sensitive for slight RI changes. In addition, these peaks are preferred to be located in the desired wavelength range. In general, the Si NP array behaves like a Fabry-Pérot cavity and light interference from the upper and bottom surfaces result in reflectance peaks. For a particular light wavelength, a NP can support single or several optical modes depending on its geometric dimensions and RI. In addition, guided light can be absorbed in the NPs depending on the material’s electronic band-gap and absorption coefficient. Complex behavior is expected when NPs support multi-modes wherein interference effects occur inside the NPs. On the other hand, the NPs should be as long as possible so the optical interaction length is increased for sensitivity; and the periodicity such that sufficiently high and clear reflection peaks can be obtained. NPs with diameters in the range of 100-150 nm (and similar periods) and high aspect ratios are difficult from a fabrication point of view. Besides this, the NP array geometry should be sufficiently open (hydrophilic) enabling liquids to fill the open spaces and/or to make a uniform surface biomolecule layer on the NP surfaces. From this point of view, previous studies have shown NPs with a diameter of ~200 nm, period 500 nm and height ~1.6 µm are suitable [28]. Here, we consider NP arrays of similar dimensions and also take into account the typical shapes of the fabricated pillars. However, in the visible wavelength range such NPs are multi-mode making it rather complicated to explain the origin of specific reflectance peaks and is notattempted here. Instead, the simulated reflectance peaks are characterized with respect to peak shifts induced by refractive index changes in the structures.
For the simulations, optical constant values of Si, SiO2 and Si3N4 were taken from Palik [34]. The geometry of the Si pillars is similar as for the fabricated structures (NIL1) and is depicted in Fig. 2(a). The addition of the specific RI layer was done by adding a homogeneous layer of thickness δ on the exposed surfaces of the Si pillar array, to investigate the induced changes in the reflectance spectrum. Similarly, the RI of the medium between the Si pillars was changed to determine corresponding ‘volume sensing’. Spectrally resolved reflectance simulated for the NP array shows clear reflection peaks in the visible/NIR spectrum (500-850 nm) in Fig. 2(b) [e.g. Figure 2(b); surrounding medium, air, n = 1.00]. Four clear peaks (Peak1-4) can be distinguished and these were used for the peak shift sensitivity characterization related to surface RI sensing and an indication for the volume sensing capability as well.
Fig. 2 Simulated total reflectance spectra for Si NP arrays: (a) schematics of the simulated periodic structure. (b) an example of the peak shift due to a change in effective RI of a 10 nm layer (SiO2 or Si3N4) on the pillar structures. (c) & (d) show the linearly fitted curves obtained from the data in (b). All the peak shift values are summarized in Table 1.
In the case of the RI sensing of the medium between the Si nanopillar array, this RI was varied between values of nmedium = 1.00-1.10, with nstep = 0.01. The peak shift (Δλ) of the reflection spectrum due to this is a result due to a change of the RI (Δn). And from this the RIS for NIL1 as nmedium sensor; with RIS = Δλ/Δn in units of nm/RIU. The peak shift results show that each separate peak has its own specific sensitivity: Peak1(528 nm)→ ~200 nm/RIU, Peak2(640 nm)→ ~100 nm/RIU, Peak3(673 nm)→ ~100 nm/RIU and Peak4(730 nm)→ ~225 nm/RIU.
For the local RI (layer) sensing, two types of layers, SiO2 (n≈1.5) and Si3N4 (n≈2), have been simulated. Layer thicknesses of 0-50 nm have been used, with a step size of 10 nm and from this, for both RI layer materials, the peak shift sensitivity has been determined in units of nm/nm of added layer. Representative examples of the peak shifts of the reflection spectra are shown in Fig. 2(b) and the peak shift data for SiO2 and Si3N4 are shown in Fig. 2(c) and Fig. 2(d), respectively. The peak shift data values for both SiO2 and Si3N4 are given in Table 1.
Table 1. Simulated reflectance peak shifts due to RI change due to the surface layer
The E-field has been simulated for the entire unit cell of the Si pillar structure using a 3D frequency-domain field and power monitor to identify the E-field mapping due to mode coupling into the Si pillars, possible coupling between the neighboring pillars and surface reflections. The E-field mappings are shown for the bare Si nanopillar array and with a homogeneous SiO2 surface layer of 50 nm. Representative results are displayed in Fig. 3 for which the wavelength related to Peak4 has been taken to show the E-field. Due to the multi-mode nature of the NPs, analysis is complicated and is briefly commented upon in Section 5.
Fig. 3 The E-field distribution in a cross-sectional view within the unit cell for the Si nanopillar array structure obtained by Lumerical FDTD simulations for the reflection peaks for Peak4 (Fig. 2(a)). (a): no coating layer on the pillar structures (at 730 nm) and (b): NPs with a SiO2 coating layer of 50 nm (at 747 nm).
4. Optical characterization
The fabricated Si NP arrays, with and without surface coatings, were characterized by spectrally resolved reflectivity measurements to investigate the sensitivity as optical RI sensor as well as to determine, using the simulated data, the “optical” thickness of non-conformal coatings. The deposited (thin) surface layers/coatings are: SiO2 (n≈1.5) and Si3N4 (n≈2). In Fig. 4(a) the characteristic reflection spectrum is given for the fabricated structure showing four clearly distinguishable reflection peaks (Peak1-4) in the visible/NIR wavelength range. The spectra were obtained using a Lambda 950 UV/Vis/NIR spectrophotometer equipped with an integrating sphere. Specular reflectance is commonly used and most practical to implement using a single source and detector. However, structures such as NP arrays can have appreciable diffuse scattering. To verify that the diffuse scattering is not dominant in our structures, both total and diffuse reflectance spectra were independently measured to determine the specular reflection spectrum. The specular reflectance shown in Fig. 4(a) indicates only a slight decrease in %R compared to the total reflectance and that the characteristic peaks are identical in both the spectra. Figure 4(b) shows an example of the peak shift data for the NP array with SiO2 and Si3N4 layers with a thickness of ~10 nm, respectively.
Fig. 4 Optical characterization of the peak shift in the reflectance spectrum for the Si NP array sample showing in (a) the total and specular reflectance spectra, in (b) reflectance spectra showing the peak shifts due to an over-layer of either SiO2 or Si3N4. The peak shift values for different coating thicknesses are summarized in Tables 2 and 3.
In Tables 2 and 3 the optical characterization peak shift data is shown for the SiO2 and Si3N4 layers, respectively, for the wavelength range between 500 and 850 nm. These results will be discussed in the next section in accordance with simulation results.
Table 2. Measured reflectance peak shifts and “optical” thickness of the surface SiO2 layer
Table 3. Measured reflectance peak shifts and “optical” thickness of the surface Si3N4 layer
5. Discussion
EM modeling and simulations and optical characterization have been performed for Si NP arrays, with a hexagonal period of ~530 nm, a height of ~1350 nm and a diameter (top-bottom) of ~200-350 nm, for optical sensing purposes. In the simulations two cases of optical sensing have been assessed: the refractive index sensitivity (RIS) by homogeneously changing the RI of the medium between the pillars and by adding a localized specific RI layer on the nanopillar array structure (surface sensing).
For the first case, the highest RIS found for the four identified peaks, see Fig. 2(b), in the range 500-850 nm is at 730 nm, Peak4, with a sensitivity of ~225/RIU. On the other hand, Peak1 (at 528 nm) also shows a relatively high RIS value of ~200 nm/RIU, which is suitable for optical sensing in the green/yellow region of the light spectrum. In the literature, values of ~500 nm/RIU (at ~800 nm; NIR region) have been reported for (nano)pillar array sensing [35]. For other types of sensing devices, much higher sensitivities (even up to RIS≈7000 nm/RIU [25]) have been reported.
A factor that plays an important role in the sensitivity of the optical sensing device is the spectral resolution of the detector. For a spectrometer the spectral resolution is determined by the values given in Eq. (1) in which RF is a resolution factor, Δλ the spectral range of the spectrometer, Wp the pixel width, n the number of pixels in the detector and Ws is the slit width [36]. The spectral resolution is also related to the FWHM of the peak of interest for the peak shift sensitivity for optical sensing.
In Eq. (2) the Detection Limit (DL) is given depending on the spectral resolution (δλ) and the RI Sensitivity (RIS). When taking a typical spectral resolution of a spectrophotometer, δλ = 0.05 nm in the visible region, this will result in a lower bound DL of ~2·10−4 RIU−1. This value shows it is not the highest reported sensitivity [25], though still sensitive enough for accurate RI optical sensing applications. Though, the actual DL is limited by the precision with which the shift can be determined due to possible experimental noises (fundamental or practical), resulting in a (slightly) worse actual DL. Factors that can improve the shift precision are, e.g., by increasing the sharpness of the peaks (Q-factors) or steepness of slopes.
For the localized surface RI sensing case, two types (SiO2 and Si3N4) of RI layers were investigated for a variety of layer thicknesses (ranging between 0 and 50 nm), by both simulations and optical characterization. For a (thin) layer addition on the pillar array structures, the effective RI of the layer (could) differ(s) from the one for (planar) bulk layers and it furthermore depends on the applied deposition technique and process conditions. Often information about the exact effective RI is desired for specific application purposes. For bio sensing applications the (average) layer thickness can be related to a bio molecule concentration and the effective RI value to the specific type of bio molecule (selectivity). By sensitive optical sensing, the effective RI of a layer can be determined for which the (average) layer thickness, the RI/thickness sensitivity per layer thickness variation and the spectral resolution of the detector are the limiting factors regarding accuracy. When the RI is known, the same principle can be used to determine the (average) layer thickness (or for bio sensing: concentration).
In Fig. 2(b) and Fig. 4(b) the spectral shifts are shown due to the addition of a (~)10 nm layer thickness, of either SiO2 or Si3N4, obtained by FDTD simulations and optical characterization, respectively. The characteristic reflection spectra for the simulation and the optical characterization show similar results for the reflection peak locations (though amplitude values differ) and both show a larger peak shift (~1.5-2x larger) for a higher RI layer material (n changing from 1.5 to 2); concluding a higher sensitivity for a higher RI material.
Figure 2(c) and Fig. 2(d) show the linearly fitted simulated peak shift data for the simulation. In Tables 1-3 all the peak shift results are given for all the indicated reflection peaks. The reflection peak wavelengths of the array differ slightly (few nm) between the SiO2 and Si3N4 cases, most likely induced by the repeated fabrication steps on the same sample. The sequential wet chemical etch removal of the SiO2 layer thickness by HF, which was performed first, the geometry of the Si pillar array structure could be slightly influenced (thinning) by the formation of a (very) thin native oxide layer on the surface before layer deposition and therefore resulting in a blue shift of the characteristic reflection spectrum after the removal by HF. For this reason, before each layer deposition, a reference reflection spectrum was taken to compensate for this blue shift. It can be observed from the peak shift data that for both RI layer cases, the reflection peak for a specific wavelength shows its own RI/layer thickness sensitivity. The peak shift sensitivity slightly decreases for the wavelength range 600-700 nm, compared to the ones in the 500-600 nm wavelength range, and increases again in the 700-850 nm wavelength range. The simulation data shows the highest peak shift sensitivity (~0.3 and ~0.9 nm/nm added layer for SiO2 and Si3N4, respectively) for Peak4(730 nm). A similar trend of the peak shifts with coatings is observed experimentally. However, there are quantitative differences. As noted in Table 2, the measured peak shifts are much larger compared to simulated values and the determined optical thickness of SiO2 is also appreciably thicker than the average value estimated by SEM. On the other hand, for Si3N4 the measured peak shifts are comparable with simulated values; consistent with this the determined optical thickness is also comparable to values estimated by SEM. These results are in reasonable agreement with process conditions for PECVD deposition of the layers – as noted earlier, the process pressures and the deposition rates are much higher for SiO2 compared to Si3N4. Thus, comparing simulations and experimental data, the thickness of (non-absorbing) surface coatings relevant for optical applications can be determined.
Validation of the approach using 3D surfaces with conformal coatings, for example obtained by atomic layer deposition (ALD) will be a subject matter for our future investigations.
For the optical characterization a Lambda 950 UV/Vis/NIR Spectrophotometer equipped with an integrating sphere was used. This type of spectrometer has a UV/Vis resolution of ≤0.05 nm and a NIR resolution of ≤0.2 nm. This results in that for the peak shift sensitivity found for SiO2 and Si3N4, the absolute thinnest measurable “average” coverage that could be measured with this setup is ~8.1·10−2 nm and ~5.2·10−2 nm, respectively. Thus optical detection of layers is much better than what can be obtained with other dimension metrology tools such as SEM.
For better clarification of the reflection spectrum and/or optical sensitivity of the Si nanopillar array structure, mappings of the E-field were performed by FDTD simulations at a resonant reflection peak wavelength for the NIL1 structure (at 730 nm; Peak4) and coated with a 50 nm SiO2 layer (at 747 nm; Peak4). Figure 3 shows a cross section of the E-field distribution. These results, in general were very complex, show coupling of different modes (e.g., HE11 and HE22) into the Si pillars due to their multi-mode nature and also due to additional effects because of their shape (being much wider at the base) and length. For similar reasons, complex interference effects occur inside the pillars. For guided modes, as it propagates absorption also occurs for above band gap light and affecting eventual light intensity from the top surface of the pillars. The type of pillar array structure in this work is a complex system for which it is very hard to distinguish all the separate mode couplings and for this reason it is not done here. Next to the pillar mode coupling, part of the light is reflected from the top of the pillars and from the bottom surface. A Fabry-Pérot effect is clearly seen in Fig. 3. The fill factor of the ‘medium’ cavity between the pillar arrays is an important parameter with relation to the overall reflection spectrum. The E-field mapping seems to show some surface sensitivity though it mostly indicates that the reflection spectrum of NIL1 is most likely due to the reflections related to the pillar fill factor. The surface sensitivity can also be derived from the data in Tables 1-3 for which the found peak shifts are sensitive but not as sensitive as for other previously reported sensitivities [1,3,25].
It can be concluded that the investigated Si NP array structure is promising for bio sensing applications, but is not amongst the highest reported surface sensitivities. However, this type of sensing is still a very useful method for sensitive (localized) RI sensing or RI layer thickness sensing using light sources and detectors easily accessible in the visible range. Such NP arrays, by appropriate surface functionalization, can also be used as a surface bio-layer sensor.
6. Conclusion
For optical sensing purposes in the visible/NIR wavelength range, electromagnetic (EM) modeling and simulations and optical characterization have been performed for a Si nanopillar array structure fabricated by nanoimprint lithography (NIL). A trade-off between practical implementation and sensitivity is made in terms of ease of fabrication and wavelength range accessible to low cost Si photodetectors. The pillar array structure has a hexagonal period of ~530 nm, a height of ~1350 nm and a diameter (top-bottom) of ~200-350 nm. Peak shifts of the resonant reflection peaks, from the characteristic reflection spectrum, have been used as optical sensing method. By either changing the effective refractive index (RI) of the medium between the pillar arrays or by adding a specific (thin) RI layer on the pillar arrays, a peak shift is observed which was used to determine either the (localized) effective RI change or determine the (average) RI layer thickness of a deposited layer. Finite-difference time-domain (FDTD) simulations have been used to determine the RI Sensitivity (RIS) by varying the RI of the medium between the pillars and from this the highest peak shift sensitivity was found to be RIS≈225 nm/RIU (at 730 nm). A lower bound detection limit (DL) was found to be ~2·10−4 RIU−1, for which the actual DL is limited by experimental noises. For the effective RI layer sensing (surface sensing) both FDTD simulations and optical characterization were performed. The additions of RI layers (~0-50 nm) of two types of materials were investigated: SiO2 (n≈1.5) and Si3N4 (n≈2). The peak shift data for the simulations and optical characterization were observed to be qualitatively similar. The NP array’s reflectivity peaks shows a sensitivity of ~0.3 and ~0.9 nm/nm (at 730 nm) of added layer SiO2 or Si3N4, respectively. For SiO2 coatings on NP arrays, the measured reflectivity peak shifts and the estimated thickness significantly differs from the simulated values, while for Si3N4 experiments and simulations show very good agreement. By comparing the simulated data, the “optical” thickness of surface coatings in nanostructured optical devices can be determined. With typical detection resolution in commercial spectrometers, a thickness (~10−2 nm) of surface layers can be optically detected, which is much better than other dimension metrology tools such as SEM. Finally, the results indicate Si nanopillar arrays are promising for use as biosensors in the visible/NIR wavelength range.
Funding
Swedish Research Council (Vetenskapsrådet; 501100004359).
Acknowledgements
The authors like to acknowledge the support from the Linné center for advanced optics and photonics (ADOPT) funded by the Swedish Research Council (VR). The authors like to acknowledge Dr. Y. Désières for useful insights.
References and links
1. X. Fan, I. M. White, S. I. Shopova, H. Zhu, J. D. Suter, and Y. Sun, “Sensitive optical biosensors for unlabeled targets: A review,” Anal. Chim. Acta 620(1-2), 8–26 (2008). [CrossRef] [PubMed]
2. M. J. Bañuls, R. Puchades, and Á. Maquieira, “Chemical surface modifications for the development of silicon-based label-free integrated optical (IO) biosensors: a review,” Anal. Chim. Acta 777, 1–16 (2013). [CrossRef] [PubMed]
3. D. Yelin, D. Oron, S. Thiberge, E. Moses, and Y. Silberberg, “Multiphoton plasmon-resonance microscopy,” Opt. Express 11(12), 1385–1391 (2003). [CrossRef] [PubMed]
4. D.-W. Duan, Y.-J. Rao, and T. Zhu, “High sensitivity gas refractometer based on all-fiber open-cavity Fabry-Perot interferometer formed by large lateral offset splicing,” J. Opt. Soc. Am. B 29(5), 912–915 (2012). [CrossRef]
5. J. Hodgkinson and R. P. Tatam, “Optical gas sensing: a review,” Meas. Sci. Technol. 24(1), 59 (2012).
6. X. Yang, A. S. P. Chang, B. Chen, C. Gu, and T. C. Bond, “High sensitivity gas sensing by Raman spectroscopy in photonic crystal fiber,” Sens. Actuat. B 176, 64–68 (2013). [CrossRef]
7. N. R. Stradiotto, H. Yamanaka, and M. V. B. Zanoni, “Electrochemical sensors: a powerful tool in analytical chemistry,” J. Braz. Chem. Soc. 14(2), 159–173 (2003). [CrossRef]
8. D. Grieshaber, R. MacKenzie, J. Vörös, and E. Reimhult, “Electrochemical biosensors – sensor principle and architectures,” Sensors (Basel) 8(3), 1400–1458 (2008). [CrossRef] [PubMed]
9. A. Janshoff, H.-J. Galla, and C. Steinem, “Piezoelectric mass-sensing devices as biosensors – an alternative to optical biosensors?” Angew. Chem. Int. Ed. Engl. 39(22), 4004–4032 (2000). [CrossRef] [PubMed]
10. S. Faegh, N. Jalili, and S. Sridhar, “A self-sensing piezoelectric microcantilever biosensor for detection of ultrasmall adsorbed masses: theory and experiments,” Sensors (Basel) 13(5), 6089–6108 (2013). [CrossRef] [PubMed]
11. K. Ramanathan and B. Danielsson, “Principles and applications of thermal biosensors,” Biosens. Bioelectron. 16(6), 417–423 (2001). [CrossRef] [PubMed]
12. M. Yakovleva, S. Bhand, and B. Danielsson, “The enzyme thermistor--a realistic biosensor concept. A critical review,” Anal. Chim. Acta 766, 1–12 (2013). [CrossRef] [PubMed]
13. P. Xu, K. Yao, J. Zheng, X. Guan, and Y. Shi, “Slotted photonic crystal nanobeam cavity with parabolic modulated width stack for refractive index sensing,” Opt. Express 21(22), 26908–26913 (2013). [CrossRef] [PubMed]
14. Y.-N. Zhang, Y. Zhao, and R.-Q. Lv, “A review for optical sensors based on photonic crystal cavities,” Sens. Actuat. A 233, 374–389 (2015). [CrossRef]
15. F. Dell’Olio and V. M. N. Passaro, “Optical sensing by optimized silicon slot waveguides,” Opt. Express 15(8), 4977–4993 (2007). [CrossRef] [PubMed]
16. C. A. Barrios, “Optical slot-waveguide based biochemical sensors,” Sensors (Basel) 9(6), 4751–4765 (2009). [CrossRef] [PubMed]
17. T. Wu, Y. Liu, Z. Yu, Y. Peng, C. Shu, and H. Ye, “The sensing characteristics of plasmonic waveguide with a ring resonator,” Opt. Express 22(7), 7669–7677 (2014). [CrossRef] [PubMed]
18. C. E. Campanella, F. De Leonardis, L. Mastronardi, P. Malara, G. Gagliardi, and V. M. N. Passaro, “Investigation of refractive index sensing based on Fano resonance in fiber Bragg grating ring resonators,” Opt. Express 23(11), 14301–14313 (2015). [CrossRef] [PubMed]
19. J. Yang, L. Jiang, S. Wang, B. Li, M. Wang, H. Xiao, Y. Lu, and H. Tsai, “High sensitivity of taper-based Mach-Zehnder interferometer embedded in a thinned optical fiber for refractive index sensing,” Appl. Opt. 50(28), 5503–5507 (2011). [CrossRef] [PubMed]
20. K. Kim and T. E. Murphy, “Porous silicon integrated Mach-Zehnder interferometer waveguide for biological and chemical sensing,” Opt. Express 21(17), 19488–19497 (2013). [CrossRef] [PubMed]
21. J. Homola, S. S. Yee, and G. Gauglitz, “Surface plasmon resonance sensors: review,” Sens. Actuat. B 54(1-2), 3–15 (1999). [CrossRef]
22. O. Saison-Francioso, G. Lévêque, A. Akjouj, Y. Pennec, B. Djafari-Rouhani, S. Szunerits, and R. Boukherroub, “Plasmonic nanoparticles array for high-sensitivity sensing: a theoretical investigation,” J. Phys. Chem. C 116(33), 17819–17827 (2012). [CrossRef]
23. R. J. Martín-Palma, M. Manso, and V. Torres-Costa, “Optical biosensors based on semiconductor nanostructures,” Sensors (Basel) 9(7), 5149–5172 (2009). [CrossRef] [PubMed]
24. Y. Ding, Y. Fan, Y. Zhang, Y. He, S. Sun, and H. Ma, “Fabrication and optical sensing properties of mesoporous silica nanorod arrays,” RSC Advances 5(110), 90659–90666 (2015). [CrossRef]
25. I. M. White and X. Fan, “On the performance quantification of resonant refractive index sensors,” Opt. Express 16(2), 1020–1028 (2008). [CrossRef] [PubMed]
26. B. Bhushan, D. R. Tokachichu, M. T. Keener, and S. C. Lee, “Morphology and adhesion of biomolecules on silicon based surfaces,” Acta Biomater. 1(3), 327–341 (2005). [CrossRef] [PubMed]
27. J. Escorihuela, M. J. Bañuls, J. García Castelló, V. Toccafondo, J. García-Rupérez, R. Puchades, and Á. Maquieira, “Chemical silicon surface modification and bioreceptor attachment to develop competitive integrated photonic biosensors,” Anal. Bioanal. Chem. 404(10), 2831–2840 (2012). [CrossRef] [PubMed]
28. B. D. Choudhury, R. Casquel, M. J. Bañuls, F. J. Sanza, M. F. Laguna, M. Holgado, R. Puchades, A. Maquieira, C. A. Barrios, and S. Anand, “Silicon nanopillar arrays with SiO2 overlayer for biosensing applications,” Opt. Mater. Express 4(7), 1345–1354 (2014). [CrossRef]
29. S. Xu and Z. L. Wang, “One-dimensional ZnO nanostructures: Solution growth and functional properties,” Nano Res. 4(11), 1013–1098 (2011). [CrossRef]
30. R. Yakimova, L. Selegard, V. Khranovskyy, R. Pearce, A. L. Spetz, and K. Uvdal, “ZnO materials and surface tailoring for biosensing,” Front. Biosci. 4(1), 254–278 (2012). [CrossRef] [PubMed]
31. P. Abgrall, V. Conedera, H. Camon, A.-M. Gue, and N.-T. Nguyen, “SU-8 as a structural material for labs-on-chips and microelectromechanical systems,” Electrophoresis 28(24), 4539–4551 (2007). [CrossRef] [PubMed]
32. M. Holgado, C. A. Barrios, F. J. Ortega, F. J. Sanza, R. Casquel, M. F. Laguna, M. J. Bañuls, D. López-Romero, R. Puchades, and A. Maquieira, “Label-free biosensing by means of periodic lattices of high aspect ratio SU-8 nano-pillars,” Biosens. Bioelectron. 25(12), 2553–2558 (2010). [CrossRef] [PubMed]
33. C. Pacholski, “Photonic crystal sensors based on porous silicon,” Sensors (Basel) 13(4), 4694–4713 (2013). [CrossRef] [PubMed]
34. E. D. Palik, Handbook of Optical Constants of Solids (Academic, 1985).
35. V. Canalejas-Tejero, A. López, R. Casquel, M. Holgado, and C. A. Barrios, “Sensitive metal layer-assisted guided-mode resonance SU8 nanopillar array for label-free optical biosensing,” Sens. Actuat. B 226, 204–210 (2016). [CrossRef]
36. BWTEK-Your Spectroscopy Partner, “Spectrometer Knowledge, Part 5: Spectral Resolution,” (BWTEK, 2016), http://bwtek.com/spectrometer-part-5-spectral-resolution/
37. Lumerical, “FDTD Solutions” (Lumerical, 2003–2016), https://www.lumerical.com/tcad-products/fdtd/
