1. Introduction

Optically resonant structures with high quality factor (Q), small mode volume, and cascadable design have emerged as a promising platform for compact, highly sensitive, and highly accurate diagnostics tools in the medical, military, food, and environmental sectors. In particular, on-chip resonant photonic structures including ring resonators [1–3] and multidimensional photonic crystal (PhC) [4–8] sensors have been extensively researched on silicon-on-insulator (SOI) substrates for refractometric-based biosensing applications. One-dimensional PhC nanobeams (NBs) and two dimensional PhC cavities in SOI have achieved the highest Q (>10⁷) and smallest mode volumes (< 1 (λ/n)³) [9,10], but typically only achieve bulk detection sensitivities (i.e., due to volumetric change in refractive index) near 100 nm/RIU. Such detection sensitivities are not competitive with many commercial refractometric-based biosensors including surface plasmon resonance sensors with sensitivities near 1400 nm/RIU. Furthermore, although the metric of the product of the sensitivity and Q is often discussed for the detection limit of sensors [11], for practical implementation, a device with ultra-high Q but modest sensitivity will not be desirable due to the potential for false positives with the reduced tolerance to small changes in ambient conditions. In an effort to improve the detection sensitivity of SOI on-chip photonic sensors while preserving the advantages of CMOS fabrication compatibility and miniaturization, slotted PhC NBs [12], air mode PhCs [13], and PhC with multi-hole defects [14] – all demonstrating the potential for stronger light-matter interaction between the guided mode and target molecules – have been considered. A significantly improved bulk detection sensitivity near 460 nm/RIU has been experimentally realized using an air mode PhC [13]. Moreover, using a high Q slotted PhC geometry, Di Falco et al. reported that a detection sensitivity exceeding 1500 nm/RIU can be achieved [15].

In this work, we take another approach to improving the detection sensitivity of on-chip PhC NB sensors by utilizing a porous silicon (PSi) substrate. PSi is a popular biosensing material platform due to its large internal surface area, allowing increased molecular capture and direct light-matter interaction [16,17]. Moreover, its rapid, tunable, and economic fabrication by electrochemical etching enables the straightforward formation of optical structures such as interferometers [18], waveguides [19], microcavities [20], and Bloch surface waves [21]. Prior work reporting PSi PhC biosensors refers to multilayered films of PSi that modify the spectrum of light that is incident out-of-plane; hence, these structures are not suitable for on-chip integration and, moreover, the out-of-plane PSi photonic crystal cavities usually have lower Q compared to on-chip resonant structures [22]. Several surface chemistry modifications can be implemented in PSi, and the label-free capture of DNA [23,24], proteins [25,26], viruses [27,28], and other relevant analytes has been demonstrated in both purified and complex media [29,30]. While PSi sensors have been demonstrated with bulk detection sensitivities near 400 nm/RIU measured by infiltrating solutions of varying refractive index [31], the surface area advantage of PSi is better represented by considering the surface detection sensitivity, which has been reported to be near 2000 nm/RIU for small molecule detection in a PSi Bloch surface/sub-surface wave sensor [21]. In contrast to the bulk detection sensitivity that considers a change in refractive index throughout the volume of the pores and surrounding media, the surface detection sensitivity considers a change in refractive index only near the pore walls and outer surface of the sensor. Hence the surface sensitivity better conveys the surface area advantage inherent in detecting the capture of molecules on the surfaces of porous sensors. We note that the comparison of detection sensitivity in porous and nonporous sensors is not perfectly straightforward. In the case of porous sensors, analyte enters the pores and changes the effective refractive index of the porous material itself, in addition to the refractive index of the surrounding medium. Especially for small molecule detection, there is a concentration factor that needs to be considered in the performance of sensors based on porous materials since an equal concentration of molecules in a given volume of solution exposed to a porous and nonporous sensor will result in a larger number of molecules captured in the porous sensor. This concentration factor significantly increases the spectral shift that results from molecular detection in porous optical biosensors. In the past few years, standard lithographic patterning as well as patterning based on selective laser-induced oxidation has been carried out on PSi planar waveguides to define PSi Mach-Zehnder interferometers [32] and ring resonators [31,33,34]. These waveguide-based PSi structures are more amenable to on-chip integration with photonic and optoelectronic circuits and with microfluidic channels compared to PSi Fabry-Perot interferometers and microcavities. Moreover, PSi optical structures have also been realized on SOI wafers, suggesting that there is a direct integration path for waveguide-based PSi sensors with CMOS control elements [35]. Relatively high Q factors near 10⁴ have been demonstrated in PSi ring resonators, facilitating their ability to experimentally realize a bulk detection sensitivity of 380 nm/RIU in a fluidic environment, which is more than two times higher than the detection sensitivity of comparable SOI ring resonators [31]. Recently, design approaches to further improve the performance of PSi ring resonators have been reported [36]. However, a key challenge that needs to be addressed for PSi ring resonator sensors is that the dynamic sensing range is limited by the free spectral range (FSR) inherent to the ring resonator, except for real-time sensing applications in which the resonance wavelength is continuously monitored. Very large spectral shifts resulting from the attachment of a monolayer of low molecular weight species, for example, cannot be uniquely quantified when measured after spotting and drying due to the overlapping of periodic resonances. Hence, the FSR of PSi rings produces an upper limit of detection and poses a significant challenge in obtaining the small molecule sensitivity of these structures. In this work, we circumvent the FSR challenge in PSi rings by employing PSi PhC NBs. A single resonance peak of the PSi PhC NB can be tracked during sensing experiments by monitoring its relative position compared to the dielectric band edge, which allows for a large dynamic sensing range for the molecules captured inside the PSi film. We report measured Q-factors between 10³ and 10⁴_, and a small molecule surface detection sensitivity near 1000 nm/RIU. Resonance shifts of the dielectric mode PSi PhC NB sensor upon exposure to small chemical molecules and 16-base nucleic acids are approximately one order of magnitude larger than those of a similarly designed SOI PhC NB sensor.

2. Design and simulations

The PhC NB is a symmetric structure consisting of a periodic array of air holes centered in a ridge waveguide, creating a cavity region surrounded by two mirror sections. In order to ensure confinement of light in the PSi PhC NB, two layers of PSi are required: a higher refractive index guiding layer and a lower refractive index cladding layer. In this work, the refractive index (porosity) of the guiding and cladding layers are 1.86 (59%) and 1.30 (81%), respectively, as experimentally determined on PSi films fabricated with the same etching conditions as the PSi PhC NBs (discussed in Section 3). By measuring both the thickness of each PSi film with scanning electron microscopy (SEM) and the reflectance spectra, the refractive index can be determined with straightforward transfer matrix analysis and the porosity can be estimated using the Bruggeman effective medium approximation. The dimensions of the PSi PhC NB are selected based on the deterministic design method previously demonstrated for achieving high Q PhC NBs in low index contrast polymeric materials [37,38]. The deterministic design approach specifies that by linearly increasing the mirror strength of lattice holes in the PhC NB from the center of the cavity outward, a Gaussian field attenuation profile is obtained for the resonance mode, which reduces scattering losses and increases the Q of the resonant structure. Accordingly, the following parameters were selected for the PSi PhC NB. The PSi NB guiding layer width and height are 1200 nm and 535 nm, respectively. The cladding layer height is 2500 nm. The periodicity, a, of the lattice holes is 550 nm. Band structure calculations were carried out with the MIT Photonic Bands eigensolver [39] to determine the relationship between mirror strength and filling fraction of the lattice holes (Fig. 1(a)); consequently, we set the radius of the lattice holes to quadratically taper from 170 nm at the outer edges of the NB (i.e. maximum mirror strength unit cells) to 220 nm for the two lattice holes at the cavity center (i.e., minimum mirror strength unit cells). Multiple maximum mirror strength unit cells and a zero length cavity were selected to minimize mode volume and maximize Q, following the deterministic design guidelines [38]. A schematic of the PSi PhC NB is shown in the upper part of Fig. 1(b). For the simulations that guide our intuition about the field distribution in the NB, we choose 10 maximum mirror strength segments at the ends of the NB and 10 taper segments in order to minimize computational time. In experiments, we fabricated and measured PSi NBs with 10 – 20 maximum mirror strength segments and 10 – 40 taper segments to improve the measured, loaded Q. In general, increasing both the number of maximum mirror strength segments and taper segments increases the Q of the NB, but also increases computation time due to the longer photon lifetime in the cavity. All simulations assumed an air ambient to mimic experimental conditions. We note that for lab-on-chip applications that enable real-time sensing, the pores of the PSi NB would be filled with analyte solution, reducing the effective refractive index change that results from molecular binding events in the pores and, hence, reducing the detection sensitivity of the PSi NB compared to what is considered in this work.

 

Fig. 1 (a) Mirror strength of different air hole filling fractions in PSi PhC NB unit cell. (b) Field profile of cavity mode in PSi PhC NB (lower) and corresponding refractive index profile (upper) where red and blue represent high and low refractive index, respectively. Electric field distribution in the cross section of the (c) PSi PhC NB and (d) SOI PhC NB. Simulations include the full cladding region, and the center of the waveguide is set to z = 0 for both the PSi and SOI PhC NBs. The guiding layers of the PhC NBs are labeled n_H, the cladding layers are labeled n_L, and the cover regions are labeled n = 1 (air).

Download Full Size | PDF

Three-dimensional finite-difference time-domain simulations were carried out to understand the field distribution in the dielectric-mode PSi PhC NBs. Given that the pore diameters are much smaller than the near-infrared resonance wavelength, the PSi layers are treated as effective media with refractive indices of 1.86 and 1.30 for the guiding and cladding layers, respectively, as specified above. Figure 1(b) shows the expected field localization in the cavity region and the Gaussian mode profile from the top view. Figure 1(c) shows the field distribution in the cross-section of the PSi PhC NB; the field is primarily localized within the guiding layer of PSi. As a comparison, we also simulated the field distribution for a SOI PhC NB with 10 mirror segments and 10 taper segments, following a design similar to one reported previously [10] that utilized the deterministic design method (SOI NB width = 500 nm, SOI NB height = 220 nm, a = 350 nm, lattice hole radius quadratically tapers from 55 – 75 nm). Figure 1(d) shows the field distribution in the cross-section of the SOI PhC NB where it is clear that the majority of the field is confined within the NB, as expected.

For sensing applications, the detection sensitivity increases as the interaction between the guided mode and target molecules increases. The porous nature of the PSi PhC NB therefore has a key advantage because small molecules can penetrate inside the NB where the electric field has the highest intensity. For SOI PhC NBs, on the other hand, target molecules can only interact with the evanescently decaying field outside the NB or with molecules attached inside the PhC NB holes. We can roughly estimate the sensitivity enhancement of the PSi PhC NB compared to the SOI PhC NB by comparing the field confinement factor in the active sensing region of the two designs. This field confinement factor, Γ_s, is a quantitative analysis of light-matter interaction and is defined as the ratio of the electric field in the sensing area divided by the total electric field, where E is the electric field, and S is the sensing area [40].

The active sensing area of the PSi PhC NB is considered to be the entire cross-section (i.e., both the PSi guiding and cladding layers), scaled by the porosity of the guiding and cladding layers, as well as the regions within 5 nm of the top and side surfaces of the PSi PhC NB. The 5 nm length scale was chosen to roughly estimate the total thickness of an effective molecular layer comprising all the molecules attached for the nucleic acid sensing experiments. We note that while this simple model takes into the PSi porosity that effectively reduces the cross-sectional surface area that is accessible for molecule attachment, it neglects the fine details of the nanoscale pore morphology. Furthermore, while the model considers molecules entirely filling the pores and attaching as monolayers on the sides of the NB, in practice, the molecular density is determined by a combination of molecule concentration, affinity, and available binding sites. Hence, the confinement factor is an upper-bound estimate. Based on the field distribution shown in Fig. 1(c), the field confinement in the active sensing region of the PSi PhC NB is calculated to be Γ_s = 53%. For the SOI PhC NB, the electric field within the guiding and cladding layers is inaccessible to the molecules, as illustrated in Fig. 1(d), and therefore the sensing area is only considered to be the evanescent field in the air cladding within 5 nm of the top and side surfaces of the SOI NB, resulting in Γ_s = 1.1%. The sensor confinement factor of the PSi PhC NB is therefore more than 40-fold larger than that of the SOI PhC NB due to the large internal surface area promoting direct light matter interaction within the material. We note that we chose not to include the lattice air holes in our calculations because the field is primarily localized in the dielectric regions and, moreover, small contributions to the sensor confinement factor due to a sensing area within the lattice air holes is expected to be similar for the PSi and SOI PhC NBs. Nevertheless, we acknowledge that a different choice of cross-section for the calculation would likely lead to slightly different estimated confinement factors for the PSi and SOI PhC NBs.

3. Fabrication

PSi PhC NBs were fabricated in a manner similar to what was reported previously [31] in order to experimentally determine their small molecule surface detection sensitivity. The dimensions of the fabricated PSi PhC NBs were consistent with those established by the deterministic design method discussed in Section 2. First, PSi waveguides were formed by electrochemical etching of p + (0.01 Ω·cm) Si (100) in a 15% hydrofluoric acid solution in ethanol and water at room temperature. A current density of 5 mA/cm² was applied for 100 s to create the guiding layer. Next, a 48 mA/cm² current density was applied for 100 s to form the cladding layer. Tuning the current density and time duration of the electrochemical etch allows control over the porosity (i.e., refractive index), pore size, and thickness of the PSi layers formed. A 1.5 mM solution of KOH in ethanol was drop cast on the films for 5 min to enlarge the pore diameters, followed by a thorough ethanol rinse. The samples were then thermally oxidized at 500°C in air for 5 min. Based on scanning electron microscopy (SEM) imaging, we estimate the average pore size of the PSi guiding and cladding layers to be 12 nm and 45 nm, respectively. NB patterns were transferred to the PSi waveguide films via electron beam lithography and reactive ion etching. ZEP 520A was spun at 2000 rpm and exposed by a JEOL JBX-9300-100kV electron beam lithography tool. Samples were then developed in xylenes for 30 s. Next, the NB pattern was etched into the guiding layer of the PSi waveguide with an Oxford Plasmalab 100 reactive ion etcher using C₄F₈/SF₆/Ar gases. Residual photoresist was removed in an N-methyl-2-pyrrolidone bath for 1 hour at 70°C. SOI PhC NBs were fabricated on SOI wafers with a 220 nm silicon device layer and a 3 μm buried oxide layer using the same oxidation conditions and lithographic processing steps as the PSi PhC NBs. The dimensions of the SOI PhC NBs are as specified in Section 2; the number of maximum mirror strength and taper segments was swept between 5 – 20 segments and 15 – 25 segments, respectively, for different SOI NB samples.

SEM images of the PSi PhC NB are shown in Fig. 2. Figure 2(a) shows the entire structure of one sample consisting of 30 mirror segments and 25 taper segments. Figure 2(b) is a close up of the mirror segments to illustrate the uniformity of holes in PSi and the different sized pores of the guiding and cladding layers. The roughness of the ridge and air holes is largely dictated by the pore size. The coupling PSi ridge waveguide is shown in the cross-sectional SEM image in Fig. 2(c) where it can be seen that the regions adjacent to the ridge have been etched down to the PSi cladding layer to provide the necessary lateral confinement.

 

Fig. 2 (a) SEM image of PSi PhC NB cavity with 30 mirror and 25 cavity segments and (b) zoom in of air holes in maximum mirror strength region. (c) Cross-sectional SEM image of the ridge waveguide used to couple light from a tapered fiber into the PSi PhC NB. The image was taken after the sample was manually cleaved.

Download Full Size | PDF

4. Functionalization protocols

Oxidized PSi and SOI PhC NBs were functionalized with 3-aminopropyltriethoxysilane (APTES) and succinimidyl 3-(2-pyridyldithio)propionate (SPDP), small linker molecules that form near-monolayers on the PSi surface to capture a thiol-modified probe DNA sequence (5′-TAG CTA TGG TCC TCG T-3′). A 4% APTES solution in methanol and deionized (DI) water was placed on the NBs for 15 min, followed by a 15 min soak in methanol to remove excess material. Next, the samples were thermally annealed for 10 min at 150 °C in air and rinsed with methanol. A 4.0 mg/mL solution of SPDP in ethanol was placed on the NBs post APTES attachment for 30 min, followed by a 30 min soak in ethanol to remove unbound molecules. A 100 μM solution of thiol-modified 16-base probe DNA in 4-(2-hydroxyenthyl)-1-piperazineethanesulfonic acid (HEPES) buffer (pH = 7.2) was reduced by tris (20carboxyethyl) phosphine disulfide reducing gel (Pierce) (TCEP) before exposure to the samples. It is critical to ensure that no TCEP remains in the DNA solution when it is exposed to the SPDP functionalized PSi and SOI samples; TCEP may cleave SPDP from the surface, precluding DNA functionalization. To separate TCEP from the DNA solution, 40 μL of the 100 μM DNA solution was mixed with 200 μL of 0.1 M TCEP (diluted in DI water from 0.5 M) for one hour and the solution was vortexed intermittently for 20 s. After one hour, 25 μL of 3 M sodium acetate solution was added to the solution containing DNA and left in the freezer (−20 °C) for 20 min. Then the solution was centrifuged at 12,000 rpm for 10 min, which formed a white DNA pellet at the bottom of the centrifuge tube. The supernatant was decanted and the pellet was washed with 200 μL of anhydrous ethanol (Decon Labs) and dried in air. Finally, 50 μL of HEPES buffer was added to the DNA pellet and vortexed. The DNA solution was drop casted on the SPDP functionalized PSi and SOI PhC NBs and the samples were incubated for one hour. The samples were then soaked in HEPES buffer for 15 min and rinsed with 2000 μL of DI water. To demonstrate specific molecular detection, the probe DNA-functionalized PSi and SOI PhC NBs were soaked in a 1 μM complementary target PNA (ACG AGG ACC ATA GCT A) in HEPES buffer for one hour followed by rinsing in HEPES buffer and then an additional 15 min soak in HEPES buffer to remove unbound species. Finally, the samples were rinsed with 2000 μL of DI water to ensure that no salts remained on the samples. Complementary PNA is chosen for this work in order to avoid the necessity to mitigate PSi corrosion effects that are observed when DNA-DNA hybridization occurs [41–43]. For control experiments to evaluate non-specific binding in the PSi and SOI PhC NBs, a non-complementary thiol-modified DNA probe molecule (5′-GTC AGC ATC AGA AAC C-3′) was used.

5. Experimental results and discussion

In preparation for transmission measurements, the PSi and SOI PhC NB samples were manually cleaved to a width of approximately 1.5 mm. Transmission measurements were carried out using a TE-polarized tunable laser (Santec TSL-510, λ = 1500 – 1630 nm) and fiber coupled setup with a photodiode receiver (Newport 2936-C) as previously described [31]. The input laser power was adjusted between −5 dBm and 5 dBm to transmit < 30 nW for all measurements. The transmission spectrum of a PSi PhC NB is shown in Fig. 3(a), with the lowest order cavity resonance located at 1555.5 nm and the dielectric band edge estimated to be at 1595 nm. Figure 3(b) reveals that this PSi PhC NB resonance has a Q of approximately 1.3 × 10³. This PSi NB was one of the samples selected for the biosensing experiments due to its desirable resonance wavelength with respect to the dynamic range of our measurement equipment. The Q of the PSi PhC NBs used in the experiments ranged from 1 × 10³ – 2 × 10³. A maximum Q of 8.9 × 10³ was observed for another PSi PhC NB with a resonance wavelength of λ₀ = 1600 nm, as shown in Fig. 3(c), which could not be used in the sensing experiments. We postulate that there are two primary factors limiting the measured Q of the PSi PhC NBs. First, non-ideal fabrication leading to slight non-uniformities between the lattice holes and roughness at the interface between the PSi guiding and cladding layers [44], as well as a limitation in minimum achievable surface roughness that arises due to the finite size of the pores, reduces the Q. Second, the design parameters selected likely do not lead to the highest achievable Q since low refractive index contrast polymer PhC NBs [37] comprising 50 ellipsoidal-shaped maximum mirror strength segments and 50 ellipsoidal-shaped taper segments have been demonstrated with Q = 3.6 × 10⁴; we believe such a Q is near the upper bound achievable value for PSi PhC NBs. The transmission spectrum of one of the SOI PhC NBs selected for biosensing experiments is shown in Fig. 4. The Q of the SOI PhC NBs used in the sensing experiments ranged from 5 × 10³ – 1 × 10⁴.

 

Fig. 3 (a) Transmission spectrum of PSi PhC NB cavity with the lowest order measurable resonance and band edge located at 1555.5 nm and 1595 nm, respectively. (b) Lorentzian curve fit (red line) of lowest order measurable resonance in (a) indicates a cavity Q of 1,311 for one of the PSi PhC NBs used in the sensing experiments. (c) A PSi PhC NB with a resonant mode near 1618.5 nm showed the highest experimentally measured Q = 8,938.

Download Full Size | PDF

 

Fig. 4 Transmission spectrum of a typical SOI PhC NB used in the biosensing experiments. The lowest order measurable resonance is located at 1533.9 nm and the band edge is near 1571 nm. Inset shows a magnified view of the lowest order resonance and a Lorentzian curve fit (red line) that suggests a cavity Q of 11,736.

Download Full Size | PDF

The surface sensitivity of the PSi PhC NBs were benchmarked based on the non-specific detection of the small chemical linker molecule, APTES, and specific detection of the 16-base target peptide nucleic acid (PNA) sequence. The magnitude of the resonance shifts for the PSi PhC NBs were compared to those of the SOI PhC NBs. Figure 5 summarizes the results of the PSi and SOI PhC NB biosensing experiments. Both complementary and non-complementary oligonucleotide pairs were tested to demonstrate specificity. Only the lowest order mode is shown in each part of the figure for clarity; Appendix Fig. 6 shows the shifts of all modes and the photonic band edge upon molecular infiltration. Although the transmission spectra for only two PSi PhC NBs and two SOI PhC NBs are shown in Fig. 5, we report the average resonance shift values for multiple samples (N ≥ 3 in all cases) that were tested. Figures 5(a) and 5(c) show the resonance shifts after the attachment of the small chemical molecules, probe DNA, and complementary PNA on a PSi PhC NB and SOI PhC NB, respectively. The resonance shift that resulted from attachment of the small APTES molecules in the PSi PhC NB was 30.7 ± 2.9 nm. Based on the refractive index change in the PSi guiding layer that results from APTES attachment (Δn = 0.03, see Appendix for details), we determine a small molecule detection sensitivity of approximately 1000 nm/RIU for the PSi PhC NBs (i.e., 30.7 nm / 0.03 = 1023 nm/RIU). For the SOI PhC NB, APTES attachment leads to a 0.38 ± 0.06 nm red-shift, which is nearly two orders of magnitude smaller than the red-shift observed for the PSi PhC NB after APTES attachment. The large sensitivity enhancement for APTES detection in the PSi NBs is expected due to the significantly increased available surface area for binding APTES molecules throughout the PSi structure compared to APTES binding only on the outer surfaces of the SOI NBs. The small size of the APTES molecules (approximately 0.8 nm) suggests that a sensitivity enhancement greater than that estimated in Section 2 for a 5 nm molecular layer should be achievable, which is consistent with the experimental results. We note that the pronounced Fabry-Perot fringes in the measured spectra of the SOI PhC NBs are likely due to a combination of reflections that arise due to the end facets of the sample and the relatively rapid tapering of the silicon ridge waveguide width between the NB holes and the cleaved end facets. Although the Fabry-Perot fringes reduce the accuracy of the Lorentzian fits to the data, it is clear that the resonance shift that results from APTES attachment on the SOI PhC NB is significantly smaller than that of the PSi PhC NB. The resonance red-shifts that result from subsequent attachment of SPDP and probe DNA molecules are likewise much larger for the PSi PhC NB compared to the SOI PhC NB. We note there is some sample-to-sample variation on the magnitude of the resonance shifts for each molecule that is attached, which suggests that the surface chemistry is not as robust as desired, possibly due to insufficient oxidation. Alternate surface chemistries based on carbon bonds have been shown to be more stable than oxidation [45]. When the complementary PNA molecules are exposed to the PSi PhC NBs, a 1.64 ± 0.48 nm resonance shift results, corresponding to a detection sensitivity of approximately 1.6 pm/nM. In the case of the SOI PhC NB, complementary PNA hybridization leads to a 0.22 ± 0.06 nm red-shift, which is nearly one order of magnitude smaller than that of the PSi PhC NB. This enhancement is in reasonable agreement with the simulated detection sensitivity enhancement in Section 2. The lower experimental result is likely due to the reduced surface density of the nucleic acid molecules compared to the simulation, which assumes a uniform refractive index change that would correspond to complete coverage of molecules within 5 nm of the pore walls. Moreover, the sensitivity enhancement in detecting nucleic acid molecules is less than that for detecting the small APTES molecules because the larger nucleic acid molecules are not able to access all of the internal pore surface due, in part, to their larger size inhibiting access to smaller pores and limiting their surface density in the pores that are accessible.

 

Fig. 5 (a) Transmission spectra of PSi PhC NB showing the detection of small chemical linker molecules (APTES and SPDP), probe DNA, and 1 μM target PNA. (b) Transmission spectra of PSi PhC NB for control experiment in which the probe DNA is not complementary to the target PNA (i.e., mismatch probe DNA). While hybridization of complementary nucleic acids leads to a resonance red-shift, exposure of mismatched sequences leads to a small blue-shift of the PSi PhC NB resonance. (c) Transmission spectra of SOI PhC NB showing the detection of small chemical linker molecules, probe DNA, and 1 μM target PNA. (d) Transmission spectra of SOI PhC NB for control experiment with mismatched probe DNA. Hybridization of complementary nucleic acids leads to a significantly larger resonance red-shift compared to exposure of mismatched sequences. Raw data with prominent Fabry-Perot fringes are shown in the dashed lines and Lorentzian fits are shown in the solid lines. The difference in the initial resonance wavelength for the two SOI PhC NBs may be due to the measurement of different modes in the two samples. Note that the wavelength range is much smaller for the SOI PhC NBs. In all parts of the figure, only the lowest order mode is shown for clarity.

Download Full Size | PDF

 

Fig. 6 Full transmission spectra of PSi PhC NB showing the detection of APTES molecules. The lowest order mode in each spectrum is highlighted by an arrow. Mode order in the PhC NB is identified based on the resonance position relative to the photonic band edge.

Download Full Size | PDF

Figures 5(b) and 5(d) show the transmission spectra for a PSi PhC NB and SOI PhC NB, respectively, exposed to mismatched DNA-PNA oligonucleotides. As expected, there is only a very small change in the resonance wavelength of the NBs when the non-complementary species is introduced. A 1.41 ± 0.66 nm resonance shift to shorter wavelengths was measured when PNA molecules were exposed to the PSi PhC NB while a 0.03 ± 0.05 nm resonance shift was measured for the SOI PhC NB sensors. In general, a resonance shift to shorter wavelengths indicates that species have been removed from the pores. We hypothesize that slight instabilities in the functionalization, as discussed earlier, lead to removal of a small fraction of the small molecules, and this effect is more pronounced in the PSi sensors. Nevertheless, it is clear that both the PSi PhC NB sensor and SOI PhC NB sensor can discriminate between complementary and non-complementary PNA molecules.

6. Conclusion

In summary, we introduced a PSi PhC NB cavity capable of detecting molecules over a large range of concentrations. The demonstration of the PSi PhC NB overcomes the sensitivity challenges of SOI sensor platforms by increasing the available light-matter interaction area where the optical mode spatially overlaps bound target molecules, assuming the molecules are small enough to be captured in the pores. We report experimental surface detection sensitivities of approximately 1000 nm/RIU and 1.6 pm/nM for small chemical molecules and 16-base nucleic acid molecules, respectively, which are approximately one order of magnitude larger than the detection sensitivities of similarly designed SOI PhC NB sensors. In addition to their high sensitivity and relatively high Q-factors (10³ – 10⁴), the versatile PSi substrates have compatibility with standard silicon fabrication procedures and reduced cost over SOI substrates, making the PSi PhC NB a promising platform to develop compact multidimensional photonic cavities for future lab-on-a-chip and sensor array devices.

Appendix

Transmission spectrum change due to APTES attachment

Figure 6 shows that all modes of the PSi PhC NB, as well as the dielectric band edge, shift to longer wavelength upon attachment of APTES molecules. Unlike ring resonators for which all resonant modes are equally spaced, the modes of a PhC NB can be independently tracked based on their location relative to the band edge. We choose to monitor the lowest order mode because, in general, the lowest order mode has the highest Q and smallest mode volume.

Refractive index change due to APTES attachment

Experimental determination of the effective refractive index change of the PSi guiding layer was carried out using two-layer, planar PSi waveguides fabricated with the identical conditions specified in Section 3 in the main text for the guiding and cladding layers. The normal incidence reflectance of these waveguides were measured using a Varian Cary 5000 spectrophotometer before and after attachment of APTES molecules. The APTES attachment protocol was the same as that described in Section 4 in the main text. As shown in Fig. 7(a), the reflectance spectrum red-shifts after APTES attachment. A fast Fourier transform (FFT) of the reflectance data as a function of wavelength over the range 600 – 2500 nm was taken using a Hamming window to determine the effective optical thickness (EOT) of the PSi, as shown in Fig. 7(b). The EOT is defined as 2nL where n is the refractive index of the material and L is the thickness of the material. Three different EOT values can be determined from the FFT data, corresponding to the PSi guiding layer, cladding layer, and the combined guiding and cladding layers. The effective refractive index of the PSi guiding layer before and after APTES attachment can be found by simply dividing the EOT by twice the guiding layer thickness, where the guiding layer thickness was determined to be 535 nm based on SEM imaging. Based on multiple measurements, we found that the average change in effective refractive index of the PSi guiding layer due to APTES attachment is Δn = 0.03.

 

Fig. 7 (a) Reflectance spectra of 2-layer, planar PSi waveguide before and after APTES surface functionalization. (b) Fourier transform of the reflectance spectra over a reduced wavelength window revealing the effective optical thickness of the PSi guiding layer (I), cladding layer (II), and combined guiding and cladding layers (III) before and after APTES attachment.

Download Full Size | PDF

Funding

Army Research Office (W911NF-09-1-0101, W911NF-15-1-0176); National Science Foundation (ECCS1407777); National Science Foundation, Research Experiences for Undergraduates Program (DMR-1263182); National Science Foundation Graduate Research Fellowship (F. O. Afzal).

Acknowledgments

Electron beam lithography and reactive ion etching were conducted at the Center for Nanophase Materials Sciences, which is a DOE Office of Science User Facility. SEM imaging was carried out in the Vanderbilt Institute of Nanoscale Science and Engineering. The authors thank Kelsey Beavers for her assistance with the biosensing experiments.

References

1. K. De Vos, I. Bartolozzi, E. Schacht, P. Bienstman, and R. Baets, “Silicon-on-Insulator microring resonator for sensitive and label-free biosensing,” Opt. Express 15(12), 7610–7615 (2007). [CrossRef]   [PubMed]  

2. M. Iqbal, M. A. Gleeson, B. Spaugh, F. Tybor, W. G. Gunn, M. Hochberg, T. Baehr-Jones, R. C. Bailey, and L. C. Gunn, “Label-free biosensor arrays based on silicon ring resonators and high-speed optical scanning instrumentation,” IEEE J Sel Top Quant 16(3), 654–661 (2010). [CrossRef]  

3. Y. Sun and X. Fan, “Optical ring resonators for biochemical and chemical sensing,” Anal. Bioanal. Chem. 399(1), 205–211 (2011). [CrossRef]   [PubMed]  

4. M. R. Lee and P. M. Fauchet, “Two-dimensional silicon photonic crystal based biosensing platform for protein detection,” Opt. Express 15(8), 4530–4535 (2007). [CrossRef]   [PubMed]  

5. W. C. Lai, S. Chakravarty, Y. Zou, and R. T. Chen, “Silicon nano-membrane based photonic crystal microcavities for high sensitivity bio-sensing,” Opt. Lett. 37(7), 1208–1210 (2012). [CrossRef]   [PubMed]  

6. F. Liang, N. Clarke, P. Patel, M. Loncar, and Q. Quan, “Scalable photonic crystal chips for high sensitivity protein detection,” Opt. Express 21(26), 32306–32312 (2013). [CrossRef]   [PubMed]  

7. J. E. Baker, R. Sriram, and B. L. Miller, “Two-dimensional photonic crystals for sensitive microscale chemical and biochemical sensing,” Lab Chip 15(4), 971–990 (2015). [CrossRef]   [PubMed]  

8. H. Inan, M. Poyraz, F. Inci, M. A. Lifson, M. Baday, B. T. Cunningham, and U. Demirci, “Photonic crystals: emerging biosensors and their promise for point-of-care applications,” Chem. Soc. Rev. 46(2), 366–388 (2017). [CrossRef]   [PubMed]  

9. Y. Akahane, T. Asano, B. S. Song, and S. Noda, “Fine-tuned high-Q photonic-crystal nanocavity,” Opt. Express 13(4), 1202–1214 (2005). [CrossRef]   [PubMed]  

10. Q. M. Quan, P. B. Deotare, and M. Loncar, “Photonic crystal nanobeam cavity strongly coupled to the feeding waveguide,” Appl. Phys. Lett. 96(20), 203102 (2010). [CrossRef]  

11. Z. Yu and S. Fan, “Extraordinarily high spectral sensitivity in refractive index sensors using multiple optical modes,” Opt. Express 19(11), 10029–10040 (2011). [CrossRef]   [PubMed]  

12. M. G. Scullion, A. Di Falco, and T. F. Krauss, “Slotted photonic crystal cavities with integrated microfluidics for biosensing applications,” Biosens. Bioelectron. 27(1), 101–105 (2011). [CrossRef]   [PubMed]  

13. S. Kim, H. M. Kim, and Y. H. Lee, “Single nanobeam optical sensor with a high Q-factor and high sensitivity,” Opt. Lett. 40(22), 5351–5354 (2015). [CrossRef]   [PubMed]  

14. C. Kang, C. T. Phare, Y. A. Vlasov, S. Assefa, and S. M. Weiss, “Photonic crystal slab sensor with enhanced surface area,” Opt. Express 18(26), 27930–27937 (2010). [CrossRef]   [PubMed]  

15. A. Di Falco, L. O’Faolain, and T. F. Krauss, “Chemical sensing in slotted photonic crystal heterostructure cavities,” Appl. Phys. Lett. 94(6), 063503 (2009). [CrossRef]  

16. Porous Silicon for Biomedical Applications. H. A. Santos. (Woodhead Publishing 2014).

17. S. Arshavsky-Graham, N. Massad-Ivanir, E. Segal, and S. Weiss, “Porous silicon-based photonic biosensors: Current status and emerging applications,” Anal. Chem. 91(1), 441–467 (2019). [CrossRef]  

18. K. P. S. Dancil, D. P. Greiner, and M. J. Sailor, “A porous silicon optical biosensor: Detection of reversible binding of IgG to a protein A-modified surface,” J. Am. Chem. Soc. 121(34), 7925–7930 (1999). [CrossRef]  

19. G. Rong, A. Najmaie, J. E. Sipe, and S. M. Weiss, “Nanoscale porous silicon waveguide for label-free DNA sensing,” Biosens. Bioelectron. 23(10), 1572–1576 (2008). [CrossRef]   [PubMed]  

20. S. Chan, S. R. Horner, P. M. Fauchet, and B. L. Miller, “Identification of Gram negative bacteria using nanoscale silicon microcavities,” J. Am. Chem. Soc. 123(47), 11797–11798 (2001). [CrossRef]   [PubMed]  

21. G. A. Rodriguez, J. D. Ryckman, Y. Jiao, and S. M. Weiss, “A size selective porous silicon grating-coupled Bloch surface and sub-surface wave biosensor,” Biosens. Bioelectron. 53, 486–493 (2014). [CrossRef]   [PubMed]  

22. C. Pacholski, “Photonic crystal sensors based on porous silicon,” Sensors (Basel) 13(4), 4694–4713 (2013). [CrossRef]   [PubMed]  

23. L. De Stefano, P. Arcari, A. Lamberti, C. Sanges, L. Rotiroti, I. Rea, and I. Rendina, “DNA optical detection based on porous silicon technology: from biosensors to biochips,” Sensors (Basel) 7(2), 214–221 (2007). [CrossRef]  

24. G. D. Francia, V. L. Ferrara, S. Manzo, and S. Chiavarini, “Towards a label-free optical porous silicon DNA sensor,” Biosens. Bioelectron. 21(4), 661–665 (2005). [CrossRef]   [PubMed]  

25. S. Mariani, L. Pino, L. M. Strambini, L. Tedeschi, and G. Barillaro, “10 000-fold improvement in protein detection using nanostructured porous silicon interferometric aptasensors,” ACS Sens. 1(12), 1471–1479 (2016). [CrossRef]  

26. V. Paeder, V. Musi, L. Hvozdara, S. Herminjard, and H. P. Herzig, “Detection of protein aggregation with a Bloch surface wave based sensor,” Sensor Actuat B 157(1), 260–264 (2011). [CrossRef]  

27. A. M. Rossi, L. Wang, V. Reipa, and T. E. Murphy, “Porous silicon biosensor for detection of viruses,” Biosens. Bioelectron. 23(5), 741–745 (2007). [CrossRef]   [PubMed]  

28. S. Chan, P. M. Fauchet, Y. Li, L. J. Rothberg, and B. L. Miller, “Porous silicon microcavities for biosensing applications,” Phys. Status Solidi, A Appl. Res. 182(1), 541–546 (2000). [CrossRef]  

29. S. Arshavsky-Graham, N. Massad-Ivanir, F. Paratore, T. Scheper, M. Bercovici, and E. Segal, “On chip protein pre-concentration for enhancing the sensitivity of porous silicon biosensors,” ACS Sens. 2(12), 1767–1773 (2017). [CrossRef]   [PubMed]  

30. L. M. Bonanno and L. A. DeLouise, “Whole blood optical biosensor,” Biosens. Bioelectron. 23(3), 444–448 (2007). [CrossRef]   [PubMed]  

31. G. A. Rodriguez, S. Hu, and S. M. Weiss, “Porous silicon ring resonator for compact, high sensitivity biosensing applications,” Opt. Express 23(6), 7111–7119 (2015). [CrossRef]   [PubMed]  

32. 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]  

33. P. Girault, N. Lorrain, J. Lemaitre, L. Poffo, M. Guendouz, I. Hardy, M. Gadonna, A. Gutierrez, L. Bodiou, and J. Charrier, “Racetrack micro-resonators based on ridge waveguides made of porous silica,” Opt. Mater. 50, 167–174 (2015). [CrossRef]  

34. R. Caroselli, S. Ponce-Alcántara, F. P. Quilez, D. M. Sánchez, L. T. Morán, A. G. Barres, L. Bellieres, H. Bandarenka, K. Girel, V. Bondarenko, and J. García-Rupérez, “Experimental study of the sensitivity of a porous silicon ring resonator sensor using continuous in-flow measurements,” Opt. Express 25(25), 31651–31659 (2017). [CrossRef]   [PubMed]  

35. H. Zhang, Z. Jia, X. Lv, J. Zhou, L. Chen, R. Liu, and J. Ma, “Porous silicon optical microcavity biosensor on silicon-on-insulator wafer for sensitive DNA detection,” Biosens. Bioelectron. 44, 89–94 (2013). [CrossRef]   [PubMed]  

36. P. Azuelos, P. Girault, N. Lorrain, Y. Dumeige, L. Bodiou, L. Poffo, M. Guendouz, M. Thual, and J. Charrier, “Optimization of porous silicon waveguide design for micro-ring resonator sensing applications,” J. Opt. 20(8), 085301 (2018). [CrossRef]  

37. Q. Quan, I. B. Burgess, S. K. Y. Tang, D. L. Floyd, and M. Loncar, “High-Q, low index-contrast polymeric photonic crystal nanobeam cavities,” Opt. Express 19(22), 22191–22197 (2011). [CrossRef]   [PubMed]  

38. Q. Quan and M. Loncar, “Deterministic design of wavelength scale, ultra-high Q photonic crystal nanobeam cavities,” Opt. Express 19(19), 18529–18542 (2011). [CrossRef]   [PubMed]  

39. S. Johnson and J. Joannopoulos, “Block-iterative frequency-domain methods for Maxwell’s equations in a planewave basis,” Opt. Express 8(3), 173–190 (2001). [CrossRef]   [PubMed]  

40. F. Dell’Olio and V. M. N. Passaro, “Optical sensing by optimized silicon slot waveguides,” Opt. Express 15(8), 4977–4993 (2007). [CrossRef]   [PubMed]  

41. C. Steinem, A. Janshoff, V. S.-Y. Lin, N. H. Völcker, and M. Reza Ghadiri, “DNA hybridization-enhanced porous silicon corrosion: mechanistic investigators and prospect for optical interferometric biosensing,” Tetrahedron 60(49), 11259–11267 (2004). [CrossRef]  

42. Y. Zhao, J. L. Lawrie, K. R. Beavers, P. E. Laibinis, and S. M. Weiss, “Effect of DNA-induced corrosion on passivated porous silicon biosensors,” ACS Appl. Mater. Interfaces 6(16), 13510–13519 (2014). [CrossRef]   [PubMed]  

43. Y. Zhao, J. L. Lawrie, P. E. Laibinis, and S. M. Weiss, “Understanding and mitigating DNA induced corrosion in porous silicon based biosensors,” Proc. SPIE 8933, 893302 (2014). [CrossRef]  

44. P. J. Reece, G. Lerondel, W. H. Zheng, and M. Gal, “Optical microcavities with subnanometer linewidths based on porous silicon,” Appl. Phys. Lett. 81(26), 4895–4897 (2002). [CrossRef]  

45. C. K. Tsang, T. L. Kelly, M. J. Sailor, and Y. Y. Li, “Highly stable porous silicon-carbon composites as label-free optical biosensors,” ACS Nano 6(12), 10546–10554 (2012). [CrossRef]   [PubMed]