Chemical Sensors Based on Photonic Structures

The chemical sensor plays an essential role in the fields of environmental conservation and monitoring, disaster and disease prevention, and industrial analysis. A typical chemical sensor is a device that transforms chemical information in a selective and reversible way, ranging from the concentration of a specific sample component to total composition analysis, into an analytically useful signal. Much research work has been performed to achieve a chemical sensor with such excellent qualities as quick response, low cost, small size, superior sensitivity, good reversibility and selectivity, and excellent detection limit. This book introduces the latest advances on chemical sensors. It consists of 15 chapters composed by the researchers active in the field of chemical sensors, and is divided into 5 sections according to the classification following the principles of signal transducer. This collection of up-to-date information and the latest research progress on chemical sensor will provide valuable references and learning materials for all those working in the field of chemical sensors.

. Schematic diagram of a photonic bio-chemical sensor.
Colored blocks named "Sensible area" and "Photonic device", represent the receptor and the transducer of the photonic biochemical sensor, respectively.
The receptor is a fundamental part of the sensor. In fact, it has to be designed in order to catalyze the specific chemical reaction or biochemical process, exhibiting a high selectivity for the analyte or the chemical specie to be detected. Generally, the receptor can be realized by covering the sensor surface (sensible area), with polymeric layers (~ nm-thin). These ones are characterized by selective receptors, able to capture and immobilize only one type of analyte in a complex chemical sample (e.g., glucose in blood, metallic particles in water, antigen/antibody in organic solution). By this way, the sensible area is the direct maker of the sensor selectivity and can be defined as a chemical prerogative.
In an intrinsic photonic bio-chemical sensor, the transducer is generally represented by an integrated optical waveguide (e.g., optical fiber, slot-waveguide, photonic crystal waveguide, photonic wire waveguide). By this way, this photonic device has to satisfy different parallel functions. The first one consists in guiding the photonic signal from the optical source (e.g., led, laser), to the sensible area. Consequently, the photonic waveguide has the rule of transducer, enhancing the interaction between the chemical/biochemical process and the optical signal. By this way, the chemical information can be properly transduced into an optical one and, finally, guided to the photo-detector to be transformed into an analytically useful electrical signal. In this context, the transducer can be considered www.intechopen.com as an engineering prerogative, because several technical solutions have to be implemented in order to maximize the confinement of the optical field in the sensible area.
The remaining blocks labeled as "signal conditioning" and "signal processing", represent the sensor front-end, generally identified by integrated digital/analog CMOS electronics. The electrical signal obtained by the conversion of the photonic one, has to be restored by amplification and filtering, then processed in order to decode the analyte (chemical) information to be recorded and displayed ("record and display").
Intrinsic photonic biosensors are generally integrated chips, characterized by small footprints (~ mm 2 , ~ m 2 ). On the contrary, in extrinsic photonic biochemical sensors the waveguide does not directly interact with the chemical/biochemical process. In fact, the photonic waveguide cannot represent the transducer because it is designed only to able the optical signal propagation. Extrinsic photonic sensors can be classified as "benchtop" systems. They are expensive and not suitable for mass-scale production. In addition, they are characterized by large dimensions and low portability.
Nowadays, integrated photonic chemical sensors represent the state-of-the-art of photonic sensing. In fact, these sensing systems exhibit ultra high performance and can be realized by standard fabrication processes available in Microelectronics industry (e.g., photolithography and ICP etching). Consequently, additional start-up costs to be invested in clean room equipments updating are not needed, and mass-scale production can represent a concrete business.
The most popular technological platform adopted for the fabrication of integrated photonic sensors is called silicon-on-insulator (SOI). This one works by placing a thin, insulating layer, such as silicon dioxide or glass, between a thin layer of silicon and the silicon substrate. In In particular, it is possible to recognize the insulating layer (SiO 2 ) characterized by a refractive index n SiO2 = 1.45, and built on the silicon substrate (bulk).
A thin silicon layer (n Si = 3.45 @ 1.55 m) is properly etched in order to realize the rib waveguide, covered by air (n air = 1). The architecture presented above exhibits a good optical transverse electric (TE) and transverse magnetic (TM) field confinement in the rib www.intechopen.com structure, directly exposed to the cover medium and characterized by the highest refractive index n Si . However, in both quasi-TE and quasi-TM modal distributions, the optical field is quasi entirely confined in the high index region, reducing the interaction with analytes or gases in the cover medium. To this purpose, other SOI-based waveguide architectures, such as the SOI photonic wire waveguide sketched in Fig. 3, can be adopted in order to improve the field confinement in the cover medium, as well as the interaction between the propagating optical field and the chemical/biochemical species, then resulting in an optimized transduction process. Fig. 3. Optical E y -field distribution in a silicon-wire waveguide (quasi-TM), op = 1.55 m.
In this specific waveguide configuration (h_box = 2 m, W = 450nm, H = 250nm), the optical transverse magnetic (TM) field is concentrated at both interfaces between the high refractive index region (silicon core) and the low refractive index ones (air at the upper side and SiO 2 at the bottom one). By this way, photonic wire waveguides represent a better choice if compared with previously analyzed rib ones. For an electromagnetic wave propagating in the z direction (Fig. 3), the major E-field component of the quasi-TM eigenmode undergoes a discontinuity at the horizontal silicon wire interfaces that, according to Maxwell's equations, is determined by the following relation: In Eq. 1 E L is the component of the E-field, evaluated in the low refractive index (n L ) region, while E H is the same component in the high refractive index (n H ) region. In particular, the greater the refractive index contrast ∆n = n H -n L , the higher the E L -field confinement. In this context, SOI technology represents a suitable solution for photonic sensing because it is possible to obtain a high refractive index contrast ∆n (n Si -n SiO2 ≈ 2 @ 1.55 m).
By using one of the two SOI-based waveguides described in the section above, it is possible to realize an intrinsic photonic biosensor. In particular, two different sensing principles characterizing the operating regime of a photonic chemical sensor can be used, i.e. surface and homogeneous sensing .
In surface sensing a change of the modal effective index n eff of the propagating optical signal is due to a change of thickness of an ultra-thin layer of selective receptor molecules which are www.intechopen.com immobilized on the functionalized waveguide surface. In particular, when the photonic sensor surface is exposed to a complex chemical solution, only target molecules will be recognized by selective receptors and will contribute to the bio-chemical process. By this way, if the thickness of the sensor surface layer was 0 before the exposure to the chemical sample, the thickness of the same layer will be greater, for example 0 + 1 , after the selective analyte adsorption. The thickness change ∆ can be transduced in an effective index change ∆n eff .
In homogenous sensing, the modal effective index change is produced by a change of cover medium refractive index, n c . For example, by assuming that the photonic sensor is initially exposed to air (n air = 1 @ 1.55 m), when a specific gas (e.g., He, CO 2 , Ar, N 2 , C 2 H 2 ) will cover the sensor surface, a sensible cover refractive index change ∆n c will be induced. In addition, the same analysis can be carried out if the sensor is initially covered by water (n water = 1.33 @ 1.55 m) and subsequently exposed to another liquid solution (e.g., NaCl). In Table 1  Integrated silicon wire waveguides define a class of photonic chemical sensors called "photonic wire evanescent field" (PWEF), because of the interaction between the evanescent optical field and the chemical/biochemical process to be detected (see Fig. 4). In both homogeneous and surface sensing it is possible to define an important performance parameter of the photonic chemical /biochemical sensor: the sensitivity.

∫∫
(4) 0 is the free space impedance (377Ω), E and H are the electric and magnetic field vectors, respectively, n c 0 is the unperturbed value of cover medium refractive index,ẑ indicates the unit vector along z direction (propagation direction) and Γ C is the confinement factor in the cover medium .
In surface sensing, shift of n eff (∆n eff ) due to a change ∆ of ad-layer thickness can be calculated by using a perturbation approach as: where n m is the molecular ad-layer refractive index and M is the region in which the ad-layer increases. Surface sensitivity S s , can be calculated by using the definition in Eq. 2.
In conclusion, there is a last consideration concerning with units of measurement of both analyzed sensitivities, S h and S s . In the first case, the performance parameter S h is dimensionless, because it is the ratio of two dimensionless physical quantities. On the contrary, surface sensitivity is generally measured in nm -1 , because the ad-layer due to analyte or molecules adsorption is characterized by a nanometer-scale thickness (2-5 nm).
Sensor limit of detection (LOD) represents another important parameter to characterize the sensor performance. It indicates the minimum resolvable signal and can be defined by taking into account the noise in the transduction signal (σ) and the sensor sensitivity (S). Generally, LOD can be calculated as the ratio σ/S. Photonic chemical sensors based on homogeneous and surface sensing are generally defined RI-based sensors, because the transduction process consists always in a modal effective index change, as previously described. However, there is another class of chemical sensors based on optical absorption detection. In particular, the absorption coefficient, usually indicated with the variable (cm -1 ), depends on the operative wavelength of the photonic signal and material (e.g., gas, solid, liquid solutions) electronic and optical properties. For example, fundamental vibrational and rotational modes associated with most inorganic and organic molecules are spectroscopically accessible within the mid-infrared range (mid-IR). By this way, the interaction between mid-IR photons and organic molecules provides particularly sharp www.intechopen.com transitions, which, despite of the wide variety of organic molecules, provide unique mid-IR absorption spectra. Therefore, it is possible to reflect the molecularly characteristic arrangement of chemical bonds within the probed molecules via the frequency position of the associated vibrations and mixed rotational-vibrational transitions. In addition to this, a lot of environmental harmful gases like carbon dioxide (CO 2 ), methane (CH 4 ) and sulfur dioxide (SO 2 ), to name a few, are characterized by absorption spectra in mid-IR, as sketched in Fig. 5. It is possible to detect specific gas species or chemical analytes with high sensitivity and selectivity, by analyzing the transmission spectra of the light intensity in a suitable wavelength range. In particular, the photonic signal intensity is linked to the gas or analyte concentration C via the Beer-Lambert law: where I is the light intensity at the end section of the whole path length (L), α linearly depends on C and I 0 is the light intensity at the initial section of the sensible area of photonic sensor. By this way, the sensor readout is an optical intensity one, thus it will be possible to register steep peaks in the transmission spectra corresponding to specific operative wavelengths, properly selected to be the absorption wavelengths of gas or molecule to be detected. Optical absorption detection allows to design photonic chemical sensors characterized by ultra low LOD (e.g., pg/mm 2 , ng/mL 1 ).
In typical bio-chemical sensors, detection of specific pathogens or proteins require transduction labeling elements, such as fluorescent dyes or radioactive isotope, in order to generate a physically useful signal from a recognition event. For example, in fluorescencebased detection the intensity of the fluorescence indicates the presence of the target molecules as well as their concentration. Although label detection exhibits high sensitivity and ultra low LOD down to a single molecule, labeling chemistry is expensive and timeconsuming, and may interfere with the function of bio-molecules. In this context, label-free photonic bio-sensing described in this paragraph, allows to preserve the natural form of target molecules and their natural interaction with selective receptors. In conclusion, labelfree photonic sensors nowadays represent a universal platform for biochemical assays.

Slot-waveguides for sensing applications
Slot waveguides represent a very interesting and promising architecture for photonic chemical and bio-chemical ultra-high performance sensing. In fact, by using slot waveguides www.intechopen.com it is possible to confine an extremely high optical field in the low refractive index region called "slot region", where the chemical solution or gas will be detected (Almeida et al., 2004;Iqbal et al., 2008). In Fig. 6, it is possible to appreciate the optical field distribution in a silicon slot waveguide and a relative geometrical definition of the slab slot architecture. Fig. 6. 2D-views of the E x -field spatial distribution (quasi-TE) solved for a Si-wire slot waveguide optimized @ 1.55 m; h = 324nm, g = 100nm, w = 180nm.
The mathematical expression of the E x -field of TM mode in a slab slot waveguide ( Fig. 6) is as follows: A 0 is an arbitrary constant, k 0 (2 / op ) is the vacuum wave number, k H is the transverse wave number in the high refractive index (n H ) region, C is the field decay coefficient in the cover medium and S is the field decay coefficient in the slot region (low RI region, n L ).
By using slot waveguides instead of PW EF or rib ones, it is possible to enhance the interaction between the propagating optical field and the chemical test sample. In fact, slotbased photonic sensors can exhibit ultra high performance (e.g., S h > 1, LOD ≈ 10 -4 RIU).
Schematic of the slab slot waveguide www.intechopen.com To this purpose, interesting results have been presented in literature. For example the slot waveguide optimized for homogeneous sensing and presented in Fig. 6, is characterized by ultra high sensitivity, S h = 1.0076 for quasi-TE propagation mode. A SOI slot waveguide, (w=135 nm, h=965 nm, g=100 nm) exhibiting ultra-high sensitivity of 1.1, has been also recently demonstrated by the authors. A value S h > 1 implies that an effective index change ∆n eff > ∆n C is induced by a cover index shift ∆n C.
Performance parameters such as sensitivity and confinement factors in cover medium and slot region, depend on waveguide geometrical parameters and materials selected for the fabrication. In fact, in Fig. 7 it is possible to note the sensitivity changes for homogenous sensing as a function of slot height "h", slot region width "g" and wire width "w", in a slot waveguide sensor. By this way, appropriate choice and design of photonic waveguide for chemical or biochemical sensing applications, have to be carried out in order to enhance all performance parameters as mentioned above (Passaro, 2009a). Nowadays, ingenious design techniques are needed to extend group IV photonics from near-IR to mid-IR wavelength range. In fact, harmful gases like carbon dioxide (CO 2 ), carbon monoxide (CO), methane (CH 4 ) and sulfure dioxide (SO 2 ) are characterized by absorption spectra in mid-IR, specially in the range 2-8 m, according to Fig. 5. To this purpose, several group IV material systems (e.g., SiGe, SiGeSn, SiGeC, GeC, SiSn) have been investigated . These alloys have been proposed for the design of slot waveguide sensors for homogeneous sensing, working at 3.39 m and 2.883 m. Interesting theoretical performances have been obtained by optimizing SOI-based slot waveguides designed for optical transparency at working op and ultra high refractive index contrast ∆n. In particular, S h > 1 has been demonstrated for two slot waveguides constituted by group IV material system layers, properly designed and stacked to form symmetrical wires, as sketched in Fig. 8  . The most important advantages of these mid-IR "slot" sensors with respect to conventional slot sensors optimized at 1.55 m, concern with higher performance, relaxed fabrication tolerances, and presence of a second-order slot mode, able to be used for sensing. Similar investigations have been also proposed about novel optical slot waveguides working at 2.883 m (Si 0.15 Ge 0.85 /Si/SiO 2 and Si 0.15 Ge 0.85 /SiO 2 ). /Ge 0.91 Sn 0.09 /Si/SiO 2 with w = 380nm, h = 520nm, t = 20nm, s = 50nm, g = 100nm; (b) Si 0.08 Ge 0.78 Sn 0.14 /Ge 0.97 C 0.03 /Si/SiO 2 with w = 390nm, h = 560nm, t = 20nm, s = 50nm, g = 100nm.
In addition, these slot waveguides exhibit ultra-high performance (S h > 1, LOD ~ 4×10 -5 ). The most important limitation that characterizes some of these intriguing alloys is the impossibility to grow the analyzed material systems selectively and directly on silicon dioxide. The effect of this limitation imposes some inevitable technological restrictions in the design of above described novel photonic sensors. For example, GeSn and SiGeSn alloys can be grown either on Si(100) or Ge/Si(100) wafers, using the SnD 4 , SiGeH 6 , Ge 2 H 6 and Si 3 H 8 hydride compounds as a source of Sn, Ge and Si constituent atoms, respectively. However, the Si-Ge-Sn class of materials, including all alloys mentioned in this section, constitutes a new paradigm in the integration of Si based electronics with optical components on a single chip. In fact, all fabrication techniques needed for new group IV alloys, are perfectly compatible with CMOS standard processes and facilities.

Fabrication tolerances
Slot waveguides and PWEF sensors presented until now, have been represented assuming an ideal geometry. In fact, vertical sidewalls ( Fig. 4 a-b, Fig. 6, Fig. 8 a-b) are very difficult to be obtained by the state-of-the-art etching processes (e.g., inductively coupled plasma, ICP). Thus, deviations from ideal case have to be considered and one of the most important parameters quantifying this effect is the tilting angle , as shown in Fig. 9. By this way, it is possible to distinguish between vertical ( =0°) and non vertical ( ≠0°) sidewalls. All well known performance parameters (S h , Γ C , Γ S ) have to be evaluated for different tilting angles ° in the range 0°-10°. To this purpose, interesting results have been plotted in Fig. 10.
www.intechopen.com All sensitivity values corresponding to = 0°, are referred to optimized slot waveguides characterized by vertical sidewalls. By observing Fig. 10a and 10b, it is evident that the greater the tilting angle, the lower the sensitivity. This effect is justified by the fact that increasing values of will produce a reduction of the slot region volume. Moreover, the slot region width g is not constant along y-direction. In conclusion, for > 4°, the slot mode cannot propagate because the optical field confinement in the slot region approaches zero.
Fabrication tolerance analysis represents an important aspect to estimate the real sensor performance. By this way, the solution to the technological problems presented above, is strictly related to the optimization of the etching process for a given combination of material systems. However, standard etching processes can assure tilting angles within 1°-2°. Generally, slot waveguides designed to operate in mid-IR wavelength range, exhibit greater tolerance margins because of their relaxed dimensions (Fig. 10a). To this purpose, < 3°-4° represents a good trade-off between technological constraints and device performance.

State of the art of slot waveguides designed for chemical sensing
All theoretical and experimental results demonstrate that slot waveguides are the best suitable photonic devices for sensing applications. In fact, if properly designed, they allow to concentrate a high optical field percentage in the cover medium, improving the interaction between the optical field and the chemical process, thus the limit of detection.
To this purpose, alternative solutions have been proposed in literature with the aim to enhance the efficiency and performance of photonic chemical sensors. For example, slot waveguides based on polymeric materials can ensure high performance, although characterized by low refractive index contrast ∆n (Bettotti et al., 2011). In addition, multipleslot waveguides represent an intriguing solution. By using multiple slot waveguides such as those presented in Fig. 11a-b-c, it is possible to significantly increase the sensitivity for both homogeneous (∆n eff /n c ~ 0.2) and surface sensing (∆n eff /∆ ~ 10 -4 RIU/nm) (Kargar & Lee, 2009;Sun et al. 2009  This improvement is justified by the fact that, in multiple-slot waveguides, the interaction between the optical field and the chemical test sample can be distributed in a larger area, in comparison to single slot waveguides. However, technological problems and reduced mode excitation usually limit the practical number of slots to three. Another slot waveguide architecture evolution is represented by horizontal slot waveguide, as sketched in Fig. 12a. By this way, the optical propagating slot mode is quasi-TM, because the E y -field is the only E-field component that undergoes discontinuities. A vertical slot fabrication can be realized by ICP etching, for example. However, etching in a very narrow region (slot region width g ~ 20-100nm) can produce large roughness in the vertical interface. By this way, in vertical slot waveguides, where the E x -field is concentrated at the interface between the high and low refractive index region, propagation losses can be very high, such as 11.6 ± 3.6 dB/cm in a single slot of 50nm (Baehr-Jones et al., 2005). This problem can be partially solved by using horizontal slot waveguides ( Fig. 12a-b). In fact, the latter can be fabricated by using thermal oxidation or deposition of different layers. By this way, the quasi TM-mode is not affected by surface roughness and propagation losses can be significantly reduced, down to 6.3 ± 0.2 dB/cm, 7.0 ± 0.2 dB/cm for a single and multiple slot waveguides, respectively (Sun et al., 2007). In Fig. 12a it is possible to see how the adsorption of target molecules on the cantilever surface, causes a small variation in the position of the www.intechopen.com functionalized cantilever. Consequently, an effective index change ∆n eff is caused by strain optic effect. Normally, the stress induced by proteins lies in a range between 3000 and 15000 ε (ε is the strain induced on the cantilever). Thus, the sensor response and the sensitivity are quite high. On the contrary, there are some aspects that limit the mass-scale production of these particular photonic sensors. First of all, the mechanical deformation, directly responsible of the sensing principle described above, has to be ensured also by very accurate technological steps, strongly influencing the sensor lifetime. Another important limitation concerns with the reliability of the sensing procedure. In fact, if the cantilever is corrupted by other molecules or particles not involved in the sensing detection, the calibration of the same sensor could be wrong, thus procedure has to be repeated more than one time.
In conclusion, both vertical and horizontal slot waveguides represent very promising photonic devices for chemical sensing applications. The high refractive index contrast between silicon and its oxide provides excellent optical confinement. By this way, it is possible to guide high optical intensity in the cover medium, preventing non linear effects in silicon, such as two-photon absorption (TPA), that can lead to high optical losses.

Advanced platforms for chemical photonic sensing
Generally, the choice of the photonic device suitable for label-free sensing applications (rib waveguide, photonic wire and slot waveguide) concerns with the possibility to ensure the best overlapping between optical field and chemical/biochemical sample. As analyzed until now, in both homogeneous and surface sensing the optical transduction leads to the effective index change ∆n eff . However, this optical variable has to be transformed into an analytically useful signal. To this purpose, several photonic integrated platforms have been proposed in literature. In particular, it is possible to distinguish between two different types of optical readout: the intensity and wavelength readout (power and wavelength interrogation, respectively). For example, photonic chemical sensors based on optical absorption are characterized by an intrinsic intensity readout, because the change of the absorption coefficient is directly linked to the output optical intensity, according to the Beer-Lambert law. By this way, it is possible to register the chemical-to-optical transduction by using a photodiode as a transducer (Fig. 1).
In this context, advanced optical platforms for chemical sensing will be investigated in this Section.

Mach-Zehnder interferometer
The operation principle of the integrated Mach-Zehnder interferometer (MZI) is illustrated in Fig. 13a. Light is introduced through a photonic waveguide (e.g., slot waveguide) in the Y-branch on the scheme left side. By this way, the input optical power I IN is split into two optical beams with an half power I IN /2 in the upper and lower arms, called "sensing" and "reference" arm, respectively. If any optical phase delay is applied to the guided mode in the sensing arm (∆φ = 0), light will be combined at the output Y-branch exhibiting an output optical power I OUT = I IN . In all different cases (∆φ ≠ 0 in sensing arm), the optical output power will be different from the input one ( In Eq. 9, L is the interaction length, thus the guiding path in which the overlapping between the optical field and the chemical test sample occurs, and λ is the operative wavelength. In Fig. 13b, it is possible to see the effect of different values of L on sensing efficiency. In fact, sensor resolution is the capability to detect the minimum effective index perturbation. By this way, it is possible to conclude that the longer the interaction length, the higher the sensor resolution. Graphically, this concept can be observed by analyzing the steepness and width of each cos 2 lobe. In conclusion, the closer the transmission spectra lobes, the smaller the detectable effective index change. Obviously, it is not convenient to realize cm-long arms because photonic chemical sensors have to be characterized by very small footprints. To this purpose, interesting technological solutions have been presented in literature. In Fig. 13c, a spiral path configuration is presented. By this way, it is possible to improve sensor performance and efficiency, without compromising the sensor dimensions. In fact, mm-long arms (~ 2-3mm) can be realized and concentrated in spiral architecture characterized by mlong diameter.
Very high integration with spiral-path configurations allows to design biosensor arrays for multiplexed real-time and label-free molecular detection. Ultra high performance can be achieved in terms of sensitivity and LOD. For example, the measured sensitivity of the 2.1 mm-long interferometer fabricated on SOI wafer with 2 m silica under 260nm of silicon, is dφ/dn C = 4930 rad/RIU, while the calculated change of polydimethylsiloxane (PDMS) refractive index upon xylene is 8.7×10 -7 RIU/ppm (Saunders et al., 2010). PDMS has been used as sensor cladding in order to enable sensing of BTEX (Benzene, Toluene, Ethylbenzene, Xylenes) volatile organic compound. In addition, a surface mass coverage of 0.3 pg/mm 2 can be detected with a silicon photonic wire MZI sensor array, with integrated SU-8 micro-fluidic channel (Densmore et al., 2009). In conclusion, the principal penalty of using MZI photonic sensors is represented by the cosine-dependent intensity function. www.intechopen.com In fact, the signal change is not easily resolvable near the maximum and minimum of the cosine function.

Surface plasmon resonance biosensors
A surface plasmon is a localized electromagnetic wave that propagates along the metaldielectric interface and exponentially decays into both media. Surface plasmons can be excited due to the resonant transfer of the incident photon energy and momentum to collectively oscillating electrons in a noble metal. The plasmon resonance condition is given by the following expression ( In Eq. 10, mr is the real part of the metal dielectric function, n C is the refractive index of light coupling glass, n S is the refractive index of sensing medium (dielectric), λ is the wavelength of incident light and k 0 is the wave number (2π/λ). The resonance condition is satisfied when (propagation constant of the light beam coupled at the angle ) is equal to SPW (propagation constant of the surface plasmon polariton). The resonant condition is expressed by Eq. 10 when θ = θ SPR . At the resonant condition, the reflected light intensity encounters a sharp dip due to optical absorption by surface plasmon wave (SPW). This dip can be easily detected by the photo-detector, as reported in the scheme proposed in Fig. 14. Bio-molecular recognition elements at the metal surface selectively capture target analytes in a chemical liquid test sample, the reflectance dip makes a shift due to the change of local refractive index near the surface. In particular, plasmonic related sensing procedure has been used for label-free detection of biological analytes at pico/femto-molar level, pesticides, immunoassays, DNA with a refractive index resolution of 1.4×10 -7 (Le et al. 2011), RNA, allergens and human blood-group identification (Jha & Sharma, 2010). For example, a SPR mid-infrared immunosensor used in a Fourier-transform infrared (FTIR) spectroscopy platform, exhibits a sensitivity of 3022 nm/RIU and a limit of detection of ~70 pg/mm 2 (DiPippo et al., 2010). Generally, the detection of the SPR change due to the adsorption of target molecules on the sensing surface, can be quantified by monitoring the resonant intensity, wavelength or angle change. Although SPR chemical sensors exhibit high sensing performance, they are characterized by some operative constrains. In particular, www.intechopen.com surface plasmon evanescent field penetrates in the sensing layer only for ~100nm. In this way, it is not possible to detect large target molecules like cells and bacteria. Another limitation concerns with the impossibility of the SPR photonic sensor, to distinguish between the refractive index surface change and the bulk solution refractive index change. Then, if the sensor is covered by a complex chemical solution (e.g., blood, bacteria-solution), it is very difficult to ensure high detection resolution. To this purpose, interesting architectures have been proposed in literature such as a novel optical structure, in which the metal layer is sandwiched by two dielectric layers with similar refractive index. In this way, both long (LSPR) and short (SSPR) surface plasmon modes can be generated at both metaldielectric interfaces, allowing to differentiate the background and surface bound refractive index change (Li et al., 2010). In conclusion, all commercialized SPR biosensors can be characterized by different coupling methods. In addition to prism coupling configuration sketched in Fig. 14, there are guided-wave, fiber optic and grating coupling configurations (Marrocco et al., 2010). Slot plasmonic waveguides have been also proposed as an integrated chemical sensor with high sensitivity (Hu et al., 2010).

Chemical sensors based on photonic crystals
Photonic crystals (PhCs) are periodic systems characterized by symmetrical separation between high dielectric and low dielectric regions. Generally, it is possible to distinguish between three different geometrical configurations as a function of the spatial periodicity. In fact, there are one-dimensional PhCs where the dielectric function periodicity is distributed only to one dimension (1D), thus bi-dimensional (2D) and three-dimensional (3D) PhCs. By using photonic crystals it is possible to guide and trap the light. In addition, photonic crystals exhibit a very large selective photonic band gap that prevents photons in the band gap from propagating in the material. By this way, it is possible to design a photonic waveguide by creating a line defect in the structure. The basic element of a photonic crystal is called "elemental" or "primary cell" and it can be represented by a line, circle, hexagonal or tetrahedral geometry, to name a few. For example, the periodical repetition of the hexagonal geometry produces a photonic crystal with a particular structure defined as "honeycomb". Photonic crystals exhibit interesting and promising features, suitable for high performance chemical sensing.

Integrated optical Bragg-grating-based chemical sensors
Photonic Bragg gratings are characterized by a periodic spatial distribution of the dielectric constant, thus of the refractive index, along the propagation direction. Consequently they are photonic crystals and exhibit a selective range of propagating wavelengths.
It is possible to realize an integrated optical sensor based on Bragg gratings in SOI technology, as sketched in Fig. 15 (Passaro et al., 2008). A Bragg grating is characterized by a periodic perturbation of the waveguide refractive index along the propagating direction z: In particular, n 0 (x,y) is the waveguide refractive index distribution, ∆n(x,y,z) represents the periodic refractive index change characterized by a period Λ (Fig. 15a). Λ represents a physical parameter that has to be designed in order to ensure a precise center wavelength of www.intechopen.com the Bragg grating spectral response (Fig. 15b), following the relation as in Eq. 12. In particular, m is the grating order, n eff is the effective index of the propagating mode and λ C is the central wavelength, previously defined.
In conclusion, the grating reflectivity can be expressed by the following expression: where k is the coupling coefficient between backward and forward propagating optical modes, L is the grating length, and are two optical parameters defined as: By this way, it is possible to explain the sensor device operative regime. In particular, all geometrical parameters (h, H, D, W, d, Λ) and the index contrast ∆n, have to be designed in order to determinate the central operative wavelength λ C . In Fig. 15b, it corresponds to 1550 nm. By observing Eq. 12, it is possible to see that λ C only depends on the effective index of the mode propagating, because the grating period Λ is fixed after the device fabrication. Consequently, any effective index change ∆n eff induced by homogeneous or surface sensing, causes a central wavelength shift, ∆λ C . In the graph reported in Fig. 15b, the sensor operative regime is characterized by λ C = 1550nm with air cover. When the sensor is exposed to aqueous solution (e.g., water), λ C = 1552.7nm. By this way, if the selected operative wavelength λ op is 1550nm, than it is possible to detect a change in the solution composition by measuring the new center wavelength shift, ∆λ C = 2.69nm. By using third-order grating based sensor devices (m = 3), it www.intechopen.com is possible to maximize the device efficiency and employ shorter length chips. In conclusion, grating sensitivity can be defined as following: In Eq. 15, n clad is the cladding refractive index and S w is the waveguide sensitivity, whose definition depends on sensing mechanism (homogeneous or surface sensing). A device sensitivity S around 120nm/RIU is practically obtainable. In conclusion, Bragg gratings analyzed above can be adopted in different waveguide architectures such fiber optics (fiber Bragg gratings, FBG). Sensitivity of 92 nm/RIU has been demonstrated for refractive index change of water in a FBG chemical sensor ).
The sensing principle described above is based on the wavelength interrogation of the Bragg grating. However, another type of interrogation can be used, based on the incident angle (°) that the propagating light forms with the Bragg grating surface. The device operation can be expressed by the following relation: Coupling wavelength λ depends on the grating period and interrogation angle. By this way, by monitoring changes in coupling angle, molecular binding at the waveguide surface can be detected and measured by using SPR sensors, as in Fig 16a-b. In particular, the reflected intensity spectrum exhibits a sharp resonance peak at the coupling wavelength. The sensing principle based on angle interrogation can be used also in surface plasmon resonance sensor characterized by a grating coupling. In particular, a refractive index resolution of 1.5×10 -6 RIU has been demonstrated by using a sensor as that sketched in Fig. 16b (Kuo & Chang, 2010).

State of the art of integrated chemical sensors based on photonic crystals
Nowadays, one and two dimensional photonic crystals represent an interesting and promising research field for sensing applications, because of their compactness and high resolution in the detection process. Recently, a photonic crystal slot waveguide has been fabricated and used as infrared (IR) absorption spectrometer for methane monitoring. The www.intechopen.com principle of IR absorption spectroscopy is based on the Beer-Lambert law, described in the previous section (Lai et al., 2011).
The photonic device shown in Fig. 17a, has been realized by using a circular geometry as primary cell. Consequently, a periodic 3D spatial distribution of the refractive index is realized with a lattice constant a. In fact, it is possible to observe that circular holes are characterized by low refractive index (n SiO2 ), while the reciprocal space represents the high refractive index region (n Si ). A slot waveguide has been realized by introducing a line defect. By this way, a high electric field intensity is localized in a low-index (n air ) 90nm wide slot. The silicon photonic crystal slot waveguide device is characterized by a total length of 300 m in order to increase the optical absorption path length. A methane concentration of 100ppm in nitrogen has been measured.
In Fig. 17b a photonic crystal air-slot cavity has been presented (Jagerska et al., 2010). The structure is similar to the previously analyzed slot waveguide. In fact, the proposed air-slot cavity has been processed on a 220nm thick silicon-on-insulator wafer with a 2 m buried oxide layer. The cavity has been realized by reducing the line defect width by 20nm, thus unchanging the photonic crystal lattice constant. Reduced-slit width results in the formation of reflective barriers for the cavity mode. By this way, only resonant wavelengths can propagate inside the photonic cavity. The sensor has been tested by exposing the chip to gases of different refractive index such as N 2 (RI = 1.00270), He (1.00032) and CO 2 (1.00407) at the operative wavelength, op = 1570nm. The refractive index change (∆n) due to the sensor exposition to different gases, causes a resonant wavelength shift (∆ ). Consequently, it is possible to detect different gas concentrations by monitoring the cavity transmission spectrum.
The sensor sensitivity can be defined as follows: www.intechopen.com In Eq. 17, λ is the resonance wavelength of the cavity, Γ is the mode field overlap and n eff the modal effective index. The high quality factor of the cavity (Q ~ 2.6×10 4 ) allows to achieve an experimental sensitivity S = 510 nm/RIU and a detection limit below 1×10 -5 RIU.
In conclusion, other intriguing solutions have been proposed such as a single-defect silicon 2D photonic crystal nano-cavity for strain sensing (Tung et al., 2011). The sensing principle is the same as the gas sensor described above, but the resonant wavelength shift is induced by geometrical longitudinal and transverse strain. The theoretical minimum detectable strain for the photonic crystal cavity has been estimated to be 8.5×10 -9 ε (8.5nε). By this way, advanced chemical sensing applications can be contemplated, based on cantilever sensing architectures.

Integrated optical directional coupler biosensors
The use of optical directional couplers for the measurement of interactions between biological molecules and the detection of target chemical and biochemical species, is an interesting research field for sensing applications. In fact, these photonic sensors are characterized by high performance and real-time monitoring of multiple outputs. The scheme of a directional coupler is presented in Fig. 18. The sensing principle is based on the super-mode theory. In Fig. 18 it is possible to see that only one waveguide (in this case the waveguide 1), is directly exposed to the chemical test sample. The other one has to be isolated by using appropriate coatings (e.g., polymers, Teflon). Generally, both waveguides have to be designed as symmetrical (e.g., identical geometries and contrast index ∆n), in order to exhibit the same propagation features. In this way, 1 and 2 , the propagation constants of the optical modes propagating in the waveguide 1 and 2, respectively, can be considered equal ( 1 = 2 ) in a "synchronous" regime, where the propagation constant mismatch is equal to zero (∆ = 0). By assuming that an optical power P 0 is introduced in the waveguide 1 at the section x = 0, optical powers (P 1 , P 2 ) at the output section (x = L) can be calculated by the following expressions: ( ) where k is the coupling coefficient and depends on the overlap between the two propagating modal fields, and L is the device length. In synchronism condition (∆ = 0), complete power transfer from the waveguide 1 to the waveguide 2 is achieved when kL = /2, 3 /2, 5 /2, etc. In this way, it will results P 2 = P 0 and P 1 = 0, at the end section x = L.
When a chemical test sample, gas or liquid solution, covers the waveguide 1, the refractive index of the cover medium will change, causing a modal effective index change of the propagating mode n eff1 . Consequently, the synchronous coupling condition will not be ensured because 1 ≠ 2 , thus ∆ ≠ 0. According to Eq. 18, P 1 will be different to zero because a complete optical power transfer now cannot occur. By monitoring the change of coupler outputs P 1 and P 2 , it is possible to detect a particular chemical target and estimate its concentration with high limit of detection. The sensitivity S can be calculated as follows: where T is the transmittance registered at the output of the reference waveguides, n clad is the chemical analyte refractive index and n clad (REF) is the reference analyte refractive index value for which the derivative is calculated. The highest sensitivity can be achieved at the point of maximum slope of the main peak of the transmittance curve. Consequently, this condition indicates the operating point of the sensor. Obviously, the overall sensitivity depends on the waveguide sensitivity, S W . In this context, the use of slot waveguides represents the best solution. To this purpose, interesting results have been obtained by using coupled slot SOI waveguides sketched in Fig. 19a. In particular, a sensitivity S = 215.29 was achieved for quasi-TE mode with a coupler length L = 400 m (Passaro et al., 2009b). The sensor sketched in Fig. 19a allows to estimate an analyte concentration in the aqueous solution. In particular, the photonic sensor integrated with a footprint of ~ 1mm 2 , can detect a minimum refractive index change of ~10 -5 and it can be adopted to estimate the glucose concentration in aqueous solution with a theoretical resolution of 0.1 g/L.
In Fig. 19b, a multi-channel directional coupler with slot waveguides is shown (McCosker & Town, 2010a). The sensing principle adopted in this device is different with respect to the www.intechopen.com previous one. In particular, it is possible to observe that there is not an isolated waveguide, because all silicon rails are simultaneously covered with the chemical solution. In this way, it is impossible to register a propagation constant mismatch change, because it will be constant. Thus, the sensing principle consists in the coupling coefficient change ∆k, as a function of the cover refractive index change ∆n clad . In particular, the coupling coefficient k has an exponential dependence on analyte refractive index (McCosker & Town, 2010b). Interesting results have been theoretically demonstrated with a 1,607 m-long device. For example, a concentration of glucose or ethanole dispersed in deionised water could be detected with sensitivities of -172 (negative T slope in Eq.19) and +155 (positive T slope), respectively.
Photonic sensors based on directional couplers, exhibit very high performance. However, it is necessary to achieve a good trade-off between sensitivity and device length. In fact, in both sensing principles as analyzed above, mm-long interaction lengths allow to appreciate transmittance changes at the device output.

Chemical sensors based on integrated optical resonant microcavities
Photonic devices exhibiting best performance for chemical sensing are those based on resonant microcavities (Passaro et al., 2007b). In fact, they allow to achieve ultra low detection limits and ultra high sensitivities due to their wavelength interrogation, poorly influenced by optical noise. Optical resonant microcavities can be designed with different paths, for example ring, race-track or spiral-path ring. Generally, a resonance structure consists in a waveguide closed into a loop and coupled with one or two input/output bus waveguides (Fig. 20a). Resonance theory imposes the propagation of precise guided stationary modes inside the resonance cavity, in particular those modes whose wavelengths satisfy the well known resonance condition: where n eff is the modal effective index, R is the ring resonator radius and m is an integer (m = 1, 2, 3,…, n), that represents the resonant order. This result has been obtained by imposing that the total round trip shift of the guided mode inside the ring resonator, must be an integer multiple of 2π. The analytical expression of the transmission spectrum of a ring resonator with a single bus (Fig. 20a), is given by the following equation: In Eq. 21, is the loss factor that gives the field attenuation after one round trip through the ring cavity waveguide (i.e., = 1 in lossless case). The cosine argument represents the round trip phase, L = 2πR is the cavity length depending on the ring radius, and t is the coupling ratio between the bus waveguide and the ring cavity.
In this context, the sensing principle consists in a resonant wavelength shift ∆λ due to the modal effective index change ∆n eff caused by the presence of the chemical analyte in the sensing area (Fig. 20b).
www.intechopen.com  The transmission spectrum shown in Fig. 20b has been calculated for a 250 m-long cavity characterized by = 0.8875, t = 0.9198. In particular, it is possible to appreciate that a modal effective index change ∆n ~ 10 -4 theoretically induces a resonant wavelength shift ∆ of about 0.1nm. As analyzed about the MZI configuration (see. Para. 3.1), in chemical sensors based on single ring resonator as sketched in Fig. 20a, the resonant cavity length assumes a key role in the photonic sensor optimization. In fact, in Fig. 21b it is possible to see that the longer the interaction length "L" between the optical field and the chemical analyte, the higher the detectable resolution.
The most important parameter that allows to appreciate the ring-based sensor resolution is the linewidth of its transmission spectrum. In fact, in Fig. 21b it is possible to see that the longer the interaction length, the sharper the resonant peak. By this way, it is possible to see that a small linewidth corresponds to a high slope of the resonant peak, resulting in a signal www.intechopen.com change easily resolvable near the maximum and minimum of the transmission function. However, very long interaction lengths will result in large device areas that are not suitable for high scale integration. To this purpose, intriguing folded spiral-path cavities have been proposed (Fig. 21a). In fact, they allow to achieve mm-long interaction length concentrated in very small footprint (~ m 2 ). The sharpness of the resonant peak in a ring resonator transmission spectrum is usually expressed in terms of its quality factor Q: In Eq. 22, λ is the resonant wavelength, ∆λ FWHM is the resonance full-width-at-half-maximum (FWHM) and n g is the ring resonator group index. It is evident that the smaller the linewidth, the higher the resonant cavity quality factor.
In conclusion, it is possible to revisit the concept of sensitivity and limit of detection (LOD): The sensitivity S (nm/RIU) depends on the waveguide sensitivity S W and other well known parameters (resonant operative wavelength and modal effective index n eff ). The limit of detection LOD (RIU), depends on the minimum detectable resonant wavelength shift ∆λ. By this way, ultra high performance for chemical, biochemical and gas sensing (S ~ 2000 nm/RIU, LOD ~ 3.8×10 -5 RIU) can be achieved by using slot waveguides .
Particular research efforts are still doing today in this field. In fact, the state-of-the-art of photonic sensors for bio-sensing applications is mainly dedicated to the optimization and characterization of novel resonant microcavities based on slot waveguides. In particular, a novel integrated SOI optic racetrack resonator has been proposed for bio-sensing applications (Malathi et al., 2010). The device is capable of distinguishing compressive and tensile stresses on a cantilever due to conformational changes of protein, Bovine Serum Albumin (BSA) and Immunoglobulin G (IgG). The change of surface stress upon adsorption of IgG is compressive, while for BSA it is tensile. The sensing principle of this specific photonic sensor configuration can be expressed by the following equation: When the bio-molecules adhere to the device surface, both effective refractive index and racetrack length change due to photo elastic effect. When the cantilever bends, one arm of the racetrack experiences a small variation in L by ∆L and n eff changes by ∆n eff , due to the strain optic effect. Bio-molecules stress on the cantilever has been simulated and the wavelength shift from the resonance has been found to be 0.3196×10 3 nm/ ε, where ε is the strain induced on the cantilever. Normally, the stress induced by proteins lies in a range between 3000 ε and 15000 ε, thus the sensor response and sensitivity obtained are quite high.
In this context, an ultrasensitive nano-mechanical photonic sensor based on horizontal slot waveguide resonator, has been also proposed on SOI platform (Barrios, 2006). The effective index variation can be read out as a resonant wavelength shift of the disk resonator, obtaining interesting results. In particular, a deflection sensitivity of 33 nm -1 , a detection limit of the slot waveguide disk resonator of 3×10 -5 nm, an estimated cantilever stress sensitivity of 1.76 nm/(mJ/m 2 ), that means that the minimum detectable surface stress is ∆σ min = 1.7×10 -5 mJ/m 2 , have been achieved.
One of the most important aspects for photonic bio-sensing concerns with temperature effects affecting final measurements. The temperature effect is directly linked to very small refractive index changes of the adopted material system. Consequently, the global modal effective index perturbation can be written as: By this way, the resonance wavelength of the ring resonator can change not only due to sensing principles (e.g., homogenous or surface sensing), but also due to temperature changes. The final result is a corrupted measure of biological analytes and molecules to be detected. Three approaches have been used for thermal noise reduction: active temperature control, a-thermal waveguide design, and temperature drift compensation by on chip referencing.
In this context, an on-chip temperature compensation in an integrated slot waveguide ring resonator refractive index sensor array has been investigated (Gylfanson et al., 2010). Experimental study has demonstrated a low temperature dependence of -16.6 pm/K while, at the same time, a large refractive index sensitivity of 240±10 nm/RIU. Furthermore, by using on chip temperature referencing, a differential temperature sensitivity of only 0.3 pm/K has been obtained, without any individual sensor calibration. This low value indicates good sensor-to-sensor repeatability, thus enabling use in highly parallel chemical assays. The detection limit has been demonstrated to be 8.8×10 -6 RIU in a 7K temperature operating windows. Another interesting solution to the problem of temperature effects in SOI wire waveguide ring resonator label-free biosensor arrays, has been proposed (Xu et al., 2010). The device is constituted by four ring resonators with a reference ring for tracking sensor temperature changes. The reference ring is protected by a 2 m-thick upper cladding layer and optically isolated from the sensing medium. By this way, real-time measurements have shown that the reference resonator resonances are linked to the temperature changes without any noticeable time delay, enabling an effective cancellation of temperature-induced shifts. A concentration of 20pM has been demonstrated by monitoring the binding between complementary IgG protein pairs. In particular, the sensor is able to detect a fluid refractive index fluctuation of ± 4×10 -6 . Better results have been achieved by using an array of 32 silicon ring resonator sensors integrated on the same platform.
With a bulk refractive index sensitivity of 7.6×10 -7 it is possible to detect a concentration of ~ 60fM, by using immobilized biotin to capture streptavidin diluted in bovine serum albumin solution (Iqbal et al., 2010).
In conclusion, a folded cavity SOI micro-ring sensor characterized by a folded spiral path geometry with a 1.2mm long ring waveguide, enclosed in a 150 m diameter sensor area, has been designed (Xu et al., 2008). The spiral cavity resonator is used to monitor the streptavidin protein binding with a detection limit of ~3 pg/mm 2 , or a total mass of ~5fg. www.intechopen.com The sensor presented above is characterized by Q-factor of 90,650 and an extinction ratio of 8.60 dB.

Resonant architectures for high performance photonic chemical sensing
Due to the increasing demand of high sensitivity and high limit of detection in biological and chemical detections, many kinds of sophisticated photonic sensors have been proposed.
In particular, the minimum detectable wavelength shift for a traditional ring sensor is given by the resonance linewidth and resolution of the optical spectra analysis (OSA). In order to achieve ultra high limit of detection, very high resolution optical spectrum analysis and very stable tunable lasers have to be usually used for the whole measurement setup, usually not convenient and very expensive. An innovative approach consists in the joint use of two resonance cavities, for example a ring resonator or a racetrack integrated inside a Fabry-Perot cavity or two ring resonators integrated in the same photonic chip (Fig. 22a). By this way, interesting ultra high performance have been demonstrated. The Fano resonance is due to complex interference present in the structure, as formed by the Fabry-Perot cavity between the reflecting elements and the ring resonator. In Fig. 22b, a typical Fano line-shape has been calculated by considering a silicon ring resonator characterized by a 30 m-long radius, integrated between two reflecting Bragg gratings with a reflection coefficient equal to 0.75. In addition, the central resonant wavelength has been imposed to be 0 = 1.535 m. Fano line-shape is characterized by a resonant peak steeper than the Lorentzian's one, that can be obtained with a standard ring resonator (Fig. 21b). By this way, the resolution of the photonic device and the limit of detection are improved. The improvement in terms of sensor performance is obtained because in Fano resonance the slope of the line-shape is greater than that obtained with Lorentzian resonance. To this purpose, a highly sensitive silicon micro-ring sensor with sharp asymmetrical resonance has been presented in literature (Yi et al., 2010). Coupled waveguides and micro-ring resonator have been fabricated using a SOI wafer which has a 1 m buffered oxide layer topped with 230nm of Si. Using a Fabry-Perot cavity characterized by a cavity length of 10mm, a quality factor Q = 3.8×10 4 and a limit of detection of ~10 -8 have been measured. www.intechopen.com The second approach allowing to ensure an ultra-sensitivity and the reduction of the detector complexity, consists in the integration of two cascaded ring resonators in the same chip. The principle used in this configuration is called "Vernier-scale" or "Vernier-effect". It consists of two wavelength scales with different periods, of which one slides the other one. The overlap between lines of two scales is used to perform the measurements.
In Fig. 23, two resonators with different optical roundtrip lengths (R 2 < R 1 ) are cascaded, so that drop signal of the first ring resonator serves as the input for the second one. Each individual ring resonator has a comb-like transmission spectrum with peaks at its resonance wavelengths, as already seen. The transmission spectrum of the cascade of two ring resonators is the product of the transmission spectra of the isolated individual resonators. Consequently, it will only exhibit peaks at wavelengths for which two resonance peaks of the respective ring resonators overlap, and the height of each of these peaks will be determined by the amount of this overlap. The sensor ring resonator will act as the sliding part of the Vernier-scale, as its evanescent field can interact with the refractive index in the environment of the sensor, where a change will cause a shift of the resonance wavelengths. The filter ring resonator is shielded from these refractive index changes by the cladding and will act as the constant part of the Vernier-scale. By this way, the whole architecture can be designed such that a small shift of the resonant wavelengths of the sensor ring resonator will result in a much larger shift of the transmission spectrum of the cascade. To this purpose, very interesting results have been demonstrated by using a silicon photonic biosensor whom architecture is the same as that sketched in Fig. 23 (Claes et al., 2010). The sensor has been tested with three aqueous solutions of NaCl with different concentrations and the operative wavelength range is from 1.52 m to 1.54 m. A sensitivity of 2169 nm/RIU has been experimentally determined in aqueous environment. In addition, the minimum detectable wavelength shift ∆ is 18pm and the LOD is 8.3×10 -6 RIU.
In conclusion, both integrated architectures analyzed above, can be designed by adopting slot waveguides in order to further improve their performance. www.intechopen.com

Conclusions
Photonic sensors represent a real revolution in chemical sensing technologies. In fact, their most important features are: • Extremely high selectivity and sensitivity; • Multi-variable and parallel processing in a chip; • Wavelength readout (noise and interference immunity); • Low-cost and high integration with front-end and support electronic systems (silicon, CMOS-compatible processing); • Real-time processing.
In addition, the silicon-on-insulator technology platform allows to integrate photonics (laser, sensing architecture, photo-detector) and electronics (CMOS data processing) on the same chip in partially monolithic form to constitute the so called Lab-on-a-chip photonic system. Moreover, the refractive index contrast of the Si/SiO 2 material system, enables record reduction in photonic device footprint with ultra-high performance and portability. Undoubtedly, photonic Lab-on-a-chip represents the innovative approach to the main mission of the modern age: the improvement of the human life quality.
Finally, a comparative The chemical sensor plays an essential role in the fields of environmental conservation and monitoring, disaster and disease prevention, and industrial analysis. A typical chemical sensor is a device that transforms chemical information in a selective and reversible way, ranging from the concentration of a specific sample component to total composition analysis, into an analytically useful signal. Much research work has been performed to achieve a chemical sensor with such excellent qualities as quick response, low cost, small size, superior sensitivity, good reversibility and selectivity, and excellent detection limit. This book introduces the latest advances on chemical sensors. It consists of 15 chapters composed by the researchers active in the field of chemical sensors, and is divided into 5 sections according to the classification following the principles of signal transducer. This collection of up-to-date information and the latest research progress on chemical sensor will provide valuable references and learning materials for all those working in the field of chemical sensors.

How to reference
In order to correctly reference this scholarly work, feel free to copy and paste the following: