Open Access
Add to BibSonomyAbstract
The presented work demonstrates the design and characterization of Si-based photonic crystal waveguides operating as an evanescent wave absorption sensor in the mid-IR range λ = 5-6 µm. The photonic crystal structure is fabricated in a Si slab upon a thin Si3N4/TEOS/Si3N4 membrane. M-line spectroscopy is used to verify the presence of guided waves. Different fillings of the photonic crystal holes have been realized to avoid sample residuals in the holes and, at the same time, to obtain spectral tuning of the structures by modification of the refractive index contrast with the photonic background. The chip displays sensitivity to fluid droplets in two-prism experiments. The output signal is quantitatively related to the fluid’s absorption coefficient thereby validating the experimental method.
© 2016 Optical Society of America
1. Introduction
Si photonics has significant potential regarding next-generation micro-optical semiconductor devices. Within the last decade numerous achievements have been realised including a Si Raman laser [1], Si optical modulators [2] and Si-Ge photodetectors [3]. The progress of group-IV photonics has not been confined solely to the laboratory, but has also been advanced by industry with Intel and IBM Corporations being two of the leaders [4]. Most breakthroughs of Si technology are however in the near infrared spectral range, for wavelengths λ = 1.3 µm and 1.55 µm, and partially in the visible range. The mid-infrared (MIR) has as yet remained underdeveloped, both in research and industry. The possibilities of implementing MIR photonics, however, are vast, specifically in development of selective chemical and bio- sensors, but also in free-space communications, laser radar transceivers, invisible-fence alarms etc. In addition, the optical transparency of Si and Ge up to wavelengths of λ = 9 µm and λ = 14 µm, respectively, allows Si photonics to be extended further into the MIR. This longer wavelength range is desirable for sensing applications (for example, in the automotive industry) since many fluids such as machine oil have distinct characteristic absorption features above λ = 5 µm.
The fundamental obstacle to developing MIR opto-electronic components thus far arises from manifold material limitations. Efficient light sources based upon group-IV materials, and capable of emitting MIR light at room temperature (RT), do not exist yet; furthermore, while quantum cascade lasers (QCL) tunable in the range λ = 3 − 5 µm have been demonstrated by Vurgaftman [5], theye are fabricated from III-V materials and are therefore incompatible with Si technology. The longer wavelength region λ = 5 – 10 µm may be yet more desirable for sensing applications due to the more spectrally distinct absorption features of many organic and inorganic chemicals/liquids. With regards to the photodetectors for this region, a breakthrough can be expected with detectors based on either ultra-small Ge quantum dots [6] or Si1-x-yGexSny alloys as shown theoretically by Sun [7].
Passive photonic components are technologically easier to fabricate than active devices but here a limitation is the small number of Si-compatible materials which show no absorption in the MIR; these include Ge, Ge-Sn alloys, Si3N4, Al2O3 and SiO2. Despite these limitations Si photonic crystals (PhC) have been demonstrated for operation at wavelengths λ = 2.9 - 3.9 µm on silicon-on-insulator (SOI) substrate [8, 9] and on sapphire [10] for the range λ = 3.45 - 3.55 µm. An interesting approach was shown in [11], where the authors fabricate PhC structures in a porous Si layer on Si substrate. There is also a trend to fabricate PhC and cavities in more exotic materials such as Ge23Sb7S70 on a CaF2 substrate amongst others [12]. All of these achievements provide a foundation for the development of MIR photonics based on photonic crystals. For the longer spectral range λ = 5-10 µm, however, these materials are either not suitable due to the high material absorption above λ = 4 µm (in Al2O3 and in glass/SOI), or they are simply not compatible with the Si-based technology (e.g. CaF2). Thus, in order to retain Si-compatible photonics while also entering the longer wavelength range, then a different material approach is necessary. In this work, we demonstrate the fabrication and characterization of two-dimensional (2D) photonic crystals (PhC) on a thin SiN/TEOS/SiN membrane via Si-based 200 mm CMOS technology (TEOS - thermally oxidized Si, further referred as SiOx). The PhCs are fabricated as “holes in a slab” and their spectral range of operation is controlled and tuned by employing different fillings of the holes rather than by changing their periodicity, , or radius, . The waveguides (WGs) are designed to be low-mode, i.e. featuring only a small number of propagating modes, in order to eliminate spurious redistribution of energy between modes during propagation, and is expected to bring clear advantages over the planar-waveguide configuration (see [13–16]).The structures display efficient guiding of light through the PhC WG at λ = 5.3 and λ = 5.5 µm along with fluid sensing in this spectral region. The entire chip contains multiple parallel PhC structures and is intended to perform as an absorption sensor for fluids in the spectral range λ = 5 − 6.2 µm.
Ideally, the device we are finally aiming at is a single-chip and would include 1) a MIR optical source, which emits an electromagnetic field propagating in 2) a Si-WG as a guided wave. The evanescent tail of the guided wave interacts with a sample (fluid), placed atop the WG, which leads to attenuation of the wave due to the partial absorption by the sample. The attenuation is the parameter used for sensing. After the interaction, the wave is guided to 3) a photodetector. Such an idealised device is shown in Fig. 1. This basic concept has been presented earlier before the requisite Si-technology was available [13–16] and the successful development of both passive and active Si/Ge components would enable such a monolithic sensing system to be realised. The incorporation of the three components onto a single chip may also be achieved via the hybrid approach, which allows for integration of non-group-IV components along with Si-compatible components on the same platform e.g. via bonding.
Fig. 1 Graphical interpretation of an idealised photonic sensor chip incorporating all the components on a single platform – a source, a waveguide, a detector and electronic components for signal readout and amplification.
2. Modelling and design
The modelling of the structures has been performed in two stages. In the first stage, spectroscopic ellipsometry [17] was implemented in order to characterize the complex dielectric function of the materials used in the fabrication. The measurements, performed on planar layers of Si3N4, SiOx and Si fabricated by liquid phase chemical vapour deposition (LPCVD) and plasma enhanced chemical vapour deposition (PECVD), were carried out using an IR VASE ellipsometer (J.A. Woollam). The measured dielectric functions of the three materials at λ = 5.3 µm and λ = 5.5 µm are presented in the table below. One important conclusion from the ellipsometry measurements is that the imaginary part of the dielectric function characterizing absorption in the material is virtually zero or close to zero (for SiOx) in the wavelength range λ = 5 - 6.2 µm, which makes the materials suitable for photonic applications. These complex dielectric functions were employed as input parameters in the photonic simulation code (see Table 1).
Table 1. The values of the complex dielectric function of the employed materials obtained by spectroscopic ellipsometry for the wavelengths λ = 5.3 and λ = 5.5 µm at which m-line spectroscopy has been carried out.
Next, photonic simulations by RSoft CAD simulation package were performed in order to optimize the photonic band gap (PhBG) [18–20], the geometries of the PhCs and to estimate the extent of the evanescent tail of the propagating field in a PhC WG formed by a linear defect (denoted W1). Two types of PhC have been considered: i) a PhC with air holes within a Si slab, and ii) a PhC with SiOx holes within a Si slab. The hole filling material (air or SiOx) primarily allows for tuning/modification of the photonic band gap (PhBG), but also ensures that no residue from the liquid sample deposited atop forms in the holes; such residue, will not only alter the PhBG, but also reduce the sensitivity of the sensor. PhCs with air holes can however be implemented for gas sensing. The index contrast modification due to the filling of the holes of the PhC slab by SiOx results in decreased confinement of the propagating field in the plane of the slab and modification of the PhBG as compared to the case of air holes. The projected band structures of the PhC slab with SiOx and air hole fillings are compared in Fig. 2(a). The band structure is calculated for a 3D symmetry-breaking slab with a semi-infinite air superstrate and SiN/SiOx substrate. First the light cone and the slab modes are calculated over the entire 2D Brillouin zone (BZ) for a defect-free PhC and then projected onto a 1D BZ with a reciprocal lattice vector corresponding to a W1 defect.
Fig. 2 (a) The projected band structure of the PhC slab on a SiN/SiOx/SiN membrane. The partial PhBG at lower frequencies includes the normalized frequency a/λ = f = 0.27 corresponding to the wavelength λ = 5.5 µm at which optical characterization was performed. The structure exhibits also a complete PhBG at higher frequencies. The dotted curves represent the guided modes in the linear defect. (b) Propagation of the guided wave along the PhC waveguide on SiN/SiOx/SiN membrane calculated by the FDTD method. Top view (top panel) and vertical cut (bottom panel) are shown. From the simulations, a light penetration depth of evanescent field on the order of couple of µm is estimated for air superstrate.
Simulations of the transmission spectrum of the PhC with the W1 defect (not shown here) demonstrate a transmission maximum spanning the normalized frequency range, along with the presence of an additional partial BG and a full band gap at higher frequencies. Figure 2(b) demonstrates the guiding of light of a wavelength λ = 5.5 µm along the PhC WG and the evanescent tail of the guided wave.
A similar, to some extent, photonic crystal system has been investigated in [21] for the telecom spectral range. The authors consider PhC fully embedded in silica aiming at improving the robustness of PhC devices as compared to those based on PhC membranes and deep study of the influence of the photonic background (air or silica) was presented. However, the PhC system demonstrated here has still air cladding (required for the operation of the device as an evanescent field sensor) and oxide filling of the holes. This fact influences the photonic characteristics like the position and the width of the band gaps, the properties of the defect modes, the light cone etc.
The photonic simulations show that a low-mode WG with a substantial evanescent tail of the guided field can be achieved for WG height of ca. 0.6 µm. This is required in order to have sufficient sensitivity of the device toward the fluid atop.
3. Experimental section
3.1. Fabrication
The samples have been fabricated using CVD, lithography and etching. The process of fabrication starts with the layer deposition performed on 8” Si wafers by means of the LPCVD technique for the nitride layer aiming at the stoichiometry Si3N4 [22] and PECVD for the SiOx. On top, a polycrystalline Si layer, which serves as a waveguiding layer, is deposited and subsequently annealed at 900°C. The PhC structures were then made using deep-UV lithography (exposure wavelength λ = 193 nm, Canon FPA-6300ES6a scanner, Canon FPA-555iZ stepper) in combination with chemical etching with HF (hydrofluoric acid). As a next step, SiOx was deposited in order to fill the holes of the PhC and chemical-mechanical polishing was applied in order to obtain flat surface, to remove the extra oxide atop and to achieve the intended thickness of the guided layer of 0.6 µm. The wafer was again annealed at 900°C to improve the crystallinity of the materials. Next, hard mask deposition and Bosch etching has been applied to the backside of the wafer in order to form the membrane. Thus, a substrate as a SiN/SiOx/SiN membrane with layer thicknesses of 140 nm/1000 nm/140 nm is achieved. The nitride is selected due to its optical transparency in the mid-infrared (see [23] and the ellipsometry measurements) and its low refractive index. We note that a thick membrane made solely from SiN was difficult to achieve due to the accumulation of stress in the layer and the associated cracking. SiOx is utilised in both the substrate and the hole filling both for its low refractive index in the MIR and relative softness for polishing.
Figure 3 shows SEM images of the fabricated structures on the chip: PhC with air holes (Fig. 3(a)) and PhC with SiOx filling (Figs. 3(b)-3(c)). The linear defect in the PhC includes a row of missing holes or holes of smaller radius to block additional frequencies allowed in the PhC WG. The estimated periodicity and radius (from SEM) of the PhC holes are a ≈1.5 µm and R ≈0.62 µm, correspondingly, which were the target parameters. It should be noted that the achieved length of the PhC waveguides were 12 mm and 15 mm, which is necessary for the evanescent sensing. This is in contrast to the e-beam technique used, for instance, in [21], which allows for short waveguides on the order of less than 2 mm or introduces stitching defects on the waveguides when attempting to fabricate long structures.
3.2. Spectroscopic characterization
The spectroscopic characterization of the PhC and their interaction with sample fluids was conducted in a series of experiments with a tunable QCL as an optical source (Daylight Solutions, 21052-MHF). The laser operates in the continuous wave regime with a power output Pmax = 140 mW and a tuning range of λ = 1776-1962 cm−1 (5.09 - 5.63 µm). The output beam from the waveguide is detected by a thermoelectrically cooled HgCdTe (MCT) detector (PVI-2TE-6, Vigo System S.A.). Measurements with both TE and TM polarizations were performed with the aid of a λ/2 wave-plate. Ge prisms were used to couple the beam into and out of the WGs.
To characterize the guiding behaviour of the structures, m-line spectroscopy in the Osterberg-Smith configuration was performed (Fig. 4(a)) [24]. It is a non-destructive method for spectroscopic characterization of photonic structures, which, in addition, allows for a separate observation of the excited guided modes, high coupling efficiencies of more than 90% and tuning the photon energy (hence the wavevector) of the incident beam in a wide range. The QCL beam enters the Ge prism from the facet AB and, after refraction, reaches the bottom facet AC (beam V3). The sample, prism, collecting lens and detector are fixed on a common rotational stage, so that only the angle of incidence θinc (and hence θ3) varies. The λ/2-wave plate allows the polarization of the laser beam to be changed from parallel to perpendicular (to the sample’s surface) and the lens focuses the beam to a waist of ca. 100 µm onto the waveguide. At certain synchronous angles, the incident wave is coupled from the prism into the waveguide and upon propagation in the WG, the wave is continuously coupled back to the prism and then detected. This coupling is registered as a peak in the transmitted signal and is a signature of a guided mode in the WG. Reflections stemming from the sample holder have been excluded by using a sample holder with a hole in it such that the incident beam is efficiently transmitted away from the detection zone. Figures 4(b) and 4(c) represent the experimental results of the m-line spectroscopy on a single PhC WG with SiOx filling for two different samples; one is a PhC slab on a membrane and the other is the PhC slab on a non-membrane (solid) substrate (see insets). Both measurements were performed at a fixed wavelength of λ = 5.5 µm with the angle of incidence changed in steps of 0.25°. Two maxima can be distinguished in the case of non-membrane sample, whereas the membrane sample shows only a single maximum. This observation indicates that only one mode (for fixed polarization) can be excited in the WG on a membrane substrate, whereas two or three modes can develop in the non-membrane case (the maximum at the highest synchronous angle corresponds to the fundamental mode).
Fig. 4 M-line experiment on Si waveguides. (a) Experimental geometry. (b) Device performance: Detected signal as a function of the angle of incidence for two polarizations of the laser beam. The PhC performance for both polarizations is explained with mode hybridization in 3D.
The behaviour of the output signal can be understood within the phenomenon of frustrated total reflection (FTR) [25,26]. The phenomenon relates tunnelling of light between two optically dense media (Ge prism, n3, and Si WG, n1) through a gap of low index of refraction (n2). In order for waveguiding to take place, the following condition must be fulfilled [25]:
In addition, for prism coupling and the FTR phenomenon, it is required thatBoth conditions are easily satisfied with the selected materials and a Ge prism (nGe = 4.1). Then, if is the wave-vector of the incident field inside the Ge prism (see the geometry in Fig. 4(a)), coupling of the incident light into the waveguide happens only when the phase matching condition is satisfied: , where is the propagation constant of the wave inside the PhC WG. In the regime of guiding, must take a value within the range [26]. From the experimental data on Fig. 4, one can measure the synchronous angle,, and hence obtain the propagation constant . From the experimental values of the , for parallel polarization, the estimated coupling coefficients are on the order of κ = 10.32 mm−1 and κ = 8.31 mm−1 for the membrane and non-membrane samples, respectively. The last expression has been adopted from the wave coupling in co-linear waveguides. The approach implemented here is comparable to the one carried out in [27] in that both are based on the phenomenon of FTR and investigate the excitation of a defect mode in a photonic crystal structure. As a basic difference, it can be pointed out that the approach demonstrated in this work investigates the excitation of the modes as a function of the angle of incidence at a fixed wavelength (hence photon energy). In contrast, the authors in Ref. 27 study the excitation of a defect mode by tuning the photon energy while keeping the angle of incidence fixed. The both approaches, thus yield different information about the behaviour of the guided modes.Next, the sensing capability of the PhC structure was examined in a two-prism experiment [25,26]. The chip contains a structure with multiple parallel PhC WGs like that shown in Fig. 3(b) with the aim of increasing of the sensor’s sensitivity. Distilled H2O was used as a sample. First, a measurement without water was performed, in which the input and output coupling is achieved via two half-prisms on each end of the WG. Efficient guiding of light in this configuration has been confirmed again via the presence of a strong output intensity maximum. Afterwards, the measurement was repeated with a water droplet atop the PhC WG. Figure 5 demonstrates the response of the PhC group to distilled H2O. The figure compares output signals from the WG without water (open symbols) and with water atop (filled symbols) at two different wavelengths, λ = 5.5 µm and λ = 5.3 µm. The signal is measured as a function of the angle of incidence in order to reach the maximum sensitivity which occurs at maximum input coupling.
Fig. 5 Results on sensing of distilled H2O: Signals without H2O (open symbols) and with H2O (filled symbols) are presented. The shift of the signal with H2O on the bottom panel is attributed to change of the coupling due to H2O penetration into the gap.
For analysis of the signal, the modified Lambert-Beer law can be employed. The guided wave propagating along a WG has a confined “core” and an evanescent tail which lies within the substrate and the superstrate. Therefore only the evanescent tail of the guided wave interacts with the sample. The penetration of the evanescent tail determines the amount of interaction between the wave and the sample atop. The Si-based PhC WG thus exhibits a number of peculiarities which introduce deviation from the Lambert-Beer law – these include the evanescent interaction of the electromagnetic wave with the substance, the strong guiding regime for wave propagation due to the large index contrast (nSi = 3.423, nair = 1, nH2O = 1.298 at λ = 5.5 µm), and the small waveguide dimensions (ca. 2 × 0.6 µm) which mean that the system is far from the ray optics limit [28] (Fig. 5(a)). The power attenuation along the WG, however, still preserves an exponential form as was demonstrated for optical fibers [29], planar [29–31] and rib [32] waveguides in the weak guiding regime and weakly absorbing samples. Assuming the same exponential power attenuation, we can write
where is the attenuation coefficient associated with absorption in the evanescent tail of the guided mode, P(x) and P0 are the power measured at the detector in the presence, or not, of an absorbing sample. The measurements with H2O samples of length x = 0.6 cm atop the chip (Figs. 5(a)-(b)) result in a power attenuation coefficient of cm−1. P0 and P(x) are then related to the absorbance byThus, at λ = 5.3 µm, the measured evanescent absorbance is while at λ = 5.5 µm, it isfor a droplet length of x = 0.6 cm. For comparison, using the values for the bulk absorbance of water listed in the NIST database, we calculate absorbance values of at λ = 5.3 µm andat λ = 5.5 µm for the experimental conditions presented here. While the measured absorbance and that calculated using the Beer-Lambert law are of the same order of magnitude, we note that deviations between the two values are common in evanescent wave studies and strongly dependent upon the strength of interaction between the silicon surface and the solution [33] (Fig. 5(a)); the larger measured absorbances compared to the calculated values (factor of 2) suggests that surface adsorption may be a significant effect and this point warrants further study at a later date.The sensing experiment has also been repeated with a second QCL operating at a slightly longer wavelength range (5.8 – 6 µm) aiming at the strong absorption peak of acetone in this range. In this case acetone droplets, formed using a volume fixed pipette, were placed upon a WG and both prisms and grating structures used as input and output couplers of the laser beam - both coupling configurations gave the same results and the spectra are in good agreement with the NIST database. Figure 6 demonstrates the absorption by acetone as a function of time and quantifies the evaporation of acetone; the earliest time spectrum shows the strongest absorption due to the acetone i.e. immediately after the droplet deposition onto the surface of the waveguide. As the acetone evaporates, the absorbance signal decreases until a minimum is reached when no liquid remains (latest time spectrum).
Fig. 6 Absorption spectrum of acetone as a function of wavelength and time. The arrow shows the chronological direction of spectra collection as evaporation of the acetone droplet proceeds.
The 20 µl acetone droplet on Si evaporates in ca. 50 seconds at room temperature. The quick evaporation hinders obtaining the angle-resolved measurement similar to Fig. 5(b) and suggests that the acetone vapours above the Si chip, before the detector, may partially contribute to the signal’s attenuation.
These observations, however, demonstrate that such a chip setup can be implemented for chemical imaging, in which the spectrum is measured as a function of time and can be uniquely associated with the volatility of the liquid.
4. Conclusions
The results presented above demonstrate that Si based MIR devices have the potential to extend the domain of integrated-Si-opto-electronics from the visible and NIR into the MIR spectral range. In particular we demonstrated that photonic crystals made of Si can be successfully implemented in the range around λ = 5.5 µm for sensing and chemical imaging applications. Tunability of the spectral range of operation can be achieved by using air- or SiOx filling of the PhC holes, though for liquid sensing applications SiOx-filling is necessary. A series of m-line experiments demonstrated guiding along Si PhC WGs and detection of liquids atop the structure. The m-line method has been applied to several different waveguides, both PhC and rib (not shown here), and successfully showed the presence of an excited mode in the waveguide. Nevertheless, one of the major difficulties of the m-line method is the mode matching. The field propagating in a Ge prism should match by its propagation constant β a guided mode in the Si WG. This is a challenge, since the field in Ge should experience a significant decrease of -vector. Integration of a MIR source on the chip could simplify the coupling and the matching problems to a big extent. At the same time, control over the lateral dimensions of the structure is vital for sensing: smaller dimensions yield an increased evanescent tail for the guided modes and stronger fluid-evanescent field interaction. This increases sensitivity, but is accompanied by the increased risk of exciting leaky modes or having strong absorption by the sample; the output signal can then fall below the detection sensitivity. The final device is contemplated as operating as a sensor for chemical or oil detection, but might also serve as a platform in chemical imaging. Before this aim can be realised, further miniaturization must occur and separation of the signal from the evanescent field in the chip, and from that of the free beam, remains a challenge; strategies for overcoming these difficulties include the integration of components as shown in Fig. 1 or signal collection through the substrate.
Acknowledgments
This work has been supported by the Austrian COMET-K2 programme of the Linz Center of Mechatronics (LCM), and was funded by the Austrian federal government and the federal state of Upper Austria. The authors express gratitude to Prof. Wolfgang Jantsch for various useful discussions and to Günter Hesser at Zentrum für Oberflächen- und Nanoanalytik (ZONA) for his SEM microscopy support. The authors express special thanks to Prof. Gus Hancock from University of Oxford for his help in various experimental and organizational issues.
References and links
1. H. Rong, A. Liu, R. Jones, O. Cohen, D. Hak, R. Nicolaescu, A. Fang, and M. Paniccia, “An all-silicon Raman laser,” Nature 433(7023), 292–294 (2005). [CrossRef] [PubMed]
2. G. T. Reed, G. Mashanovich, F. Y. Gardes, and D. J. Thomson, “Silicon optical modulators,” Nat. Photonics 4(8), 518–526 (2010). [CrossRef]
3. J. Michel, J. Liu, and L. C. Kimerling, “High performance Ge-on-Si photodetectors,” Nat. Photonics 4(8), 527–534 (2010). [CrossRef]
4. IBM news room, “Made in IBM Labs: IBM Lights Up Silicon Chips to Tackle Big Data”, http://www-03.ibm.com/press/us/en/pressrelease/39641.wss
5. I. Vurgaftman, C. L. Canedy, C. S. Kim, M. Kim, W. W. Bewley, J. R. Lindle, J. Abell, and J. R. Meyer, “Mid-infrared interband cascade lasers operating at ambient temperatures,” New J. Phys. 11(12), 125015 (2009). [CrossRef]
6. A. I. Yakimov, A. I. Nikiforov, V. A. Timofeev, A. A. Bloshkin, V. V. Kirienko, and A. V. Dvurechenskii, “Midinfrared photoresponse of Ge quantum dots on a strained Si0.65Ge0.35 layer,” Semicond. Sci. Technol. 26, 1–4 (2011).
7. G. Sun, H. H. Cheng, J. Menendez, J. B. Khurgin, and R. A. Soref, “Strain-free Ge/GeSiSn quantum cascade lasers based on L-valley intersubband transitions,” Appl. Phys. Lett. 90(25), 251105 (2007). [CrossRef]
8. C. Reimer, M. Nedeljkovic, D. J. M. Stothard, M. O. S. Esnault, C. Reardon, L. O’Faolain, M. Dunn, G. Z. Mashanovich, and T. F. Krauss, “Mid-infrared photonic crystal waveguides in silicon,” Opt. Express 20(28), 29361–29368 (2012). [CrossRef] [PubMed]
9. R. Shankar, R. Leijssen, I. Bulu, and M. Lončar, “Mid-infrared photonic crystal cavities in silicon,” Opt. Express 19(6), 5579–5586 (2011). [CrossRef] [PubMed]
10. Y. Zou, S. Chakravarty, P. Wray, and R. T. Chen, “Experimental demonstration of propagation characteristics of mid-infrared photonic crystal waveguides in silicon-on-sapphire,” Opt. Express 23(5), 6965–6975 (2015). [CrossRef] [PubMed]
11. Z. Dang, A. Banas, S. Azimi, J. Song, M. Breese, Y. Yao, S. P. Turaga, G. Recio-Sánchez, A. Bettiol, and J. V. Kan, “Silicon and porous silicon mid-infrared photonic crystals,” Appl. Phys. A Mater. Sci. Process. 112(3), 517–523 (2013). [CrossRef]
12. H. Lin, L. Li, F. Deng, C. Ni, S. Danto, J. D. Musgraves, K. Richardson, and J. Hu, “Demonstration of mid-infrared waveguide photonic crystal cavities,” Opt. Lett. 38(15), 2779–2782 (2013). [CrossRef] [PubMed]
13. J. Kasberger, Infrared Absorption Sensor,” PhD Thesis, Johannes Kepler University, Austria (2009).
14. J. Kasberger, T. Fromherz, A. Saeed, and B. Jakoby, “Design and experimental verification with deteriorated lubrication oil,” Vib. Spectrosc. 56(2), 129–135 (2011). [CrossRef]
15. G. Stewart and B. Culshaw, “Optical waveguide modelling and design for evanescent field chemical sensors,” Opt. Quantum Electron. 26(3), S249–S259 (1994). [CrossRef]
16. T. Okamoto, M. Yamamoto, and I. Yamaguchi, “Optical waveguide absorption sensor using a single coupling prism,” J. Opt. Soc. Am. A 17(10), 1880–1886 (2000). [CrossRef] [PubMed]
17. H. Fujiwara, Spectroscopic Ellipsometry Principles and Applications (John Wiley & Sons, 2007).
18. J. D. Joannopoulos, R. D. Meade, and J. N. Winn, Photonic Crystals Moulding the Flow of Light (Princeton University, 1995).
19. A. Birner, R. B. Wehrspohn, U. M. Gösele, and K. Busch, “Silicon-based photonic crystals,” Adv. Mater. 13(6), 377–388 (2001). [CrossRef]
20. K. Sakoda, Optical Properties of Photonic Crystals (Springer-Verlag, 2005).
21. T. P. White, L. O’Faolain, J. Li, L. C. Andreani, and T. F. Krauss, “Silica-embedded silicon photonic crystal waveguides,” Opt. Express 16(21), 17076–17081 (2008). [CrossRef] [PubMed]
22. H. G. Tompkins, A User’s Guide to Ellipsometry (Dover, 2006).
23. R. Soref, “Mid-infrared photonics in silicon and germanium,” Nat. Photonics 4(8), 495–497 (2010). [CrossRef]
24. H. Osterberg and L. W. Smith, “Transmission of optical energy along surfaces,” J. Opt. Soc. Am. 54(9), 1078 (1964). [CrossRef]
25. P. K. Tien and R. Ulrich, “Theory of prism-film coupler and thin-film light guides,” J. Opt. Soc. Am. 60(10), 1325–1337 (1970). [CrossRef]
26. R. Ulrich, “Theory of prism-film coupler by plane-wave analysis,” J. Opt. Soc. Am. 60(10), 1337–1350 (1970). [CrossRef]
27. M. Galli, D. Bajoni, M. Patrini, G. Guizzetti, D. Gerace, L. C. Andreani, M. Belotti, and Y. Chen, “Single-mode versus multimode behavior in silicon photonic crystal waveguides measured by attenuated total reflectance,” Phys. Rev. B 72(12), 125322 (2005). [CrossRef]
28. A. Yariv and P. Yeh, Optical Waves in Crystals (J. Weley & Sons, 1983).
29. A. W. Snyder and J. D. Love, Optical Waveguide Theory (Chapman and Hall, 1983).
30. A. Reisinger, “Characteristics of optical guided modes in lossy waveguides,” Appl. Opt. 12(5), 1015–1025 (1973). [CrossRef] [PubMed]
31. J. N. Polky and J. H. Harris, “Absorption from thin-film waveguides,” J. Opt. Soc. Am. 62(9), 1081–1087 (1972). [CrossRef]
32. R. E. Collin, Foundations for Microwave Engineering (McGraw-Hill, 1966).
33. V. Ruddy, B. D. MacCraith, and J. A. Murphy, “Evanescent wave absorption spectroscopy using multimode fibers,” J. Appl. Phys. 67(10), 6070–6074 (1990). [CrossRef]
