WO2018051145A2

WO2018051145A2 - Integrated broad-band young interferometers for simultaneous dual polarization bio-chemical sensing through amplified fringe packet shifts

Info

Publication number: WO2018051145A2
Application number: PCT/GR2017/000053
Authority: WO
Inventors: Konstantinos Misiakos; Eleni Makarona; Ionnis RAPTIS; Sotirios Kakabakos; Panagiota Petrou; Alexandros SALAPATAS; Eleftheria STAVRA
Original assignee: National Centre For Scientific Research "Demokritos" (Ncsr "Demokritos")
Priority date: 2016-09-14
Filing date: 2017-09-11
Publication date: 2018-03-22
Also published as: GR20160100477A; WO2018051145A3

Abstract

The photonic chip integrates one or more Young interferometers that accept broad-band light at their input waveguide while their output waveguides couple their light to a single imaging array either through air or through on-chip slab waveguides. The interferometers are interpolated between the input and output waveguides. The sensing window is exposed to the analyte solutions. On this window the recognition molecules have previously being immobilized. Upon analyte binding an adlayer grows on the sensing window. Depending on emitted light polarization, interference occurs at specific solid angles so that two packets of sinusoidal fringes are formed for either polarization: One for the TE and one for the TM polarization. The two different angles result from the different dispersion relationship of the phase function on the wavenumber for the two polarizations. With the development of adlayers the fringes shift in each packet following the on- chip phase change for each polarization. In addition to the fringe shift and in the same direction, each packet moves solidly as a result of the adlayer induced changes on the phase function slope against the wavenumber. In fact, the packets move faster than the on-chip phase change thus providing an amplified optical signal. The set-up incorporates the photonic chip with the broad¬ band Young interferometer and the input and output waveguides, broad-band light sources, external or on-chip integrated, and an imaging array, external or on-chip integrated. It also includes fluidic structures for the supply of reagents and the samples. Fourier transform and other techniques are employed to calculate the phase change from the fringe shift and packet motion. The packet motion and phase shifts provide a measure of the adlayer growth due to binding reactions.

Description

INTEGRATED BROAD-BAND YOUNG INTERFEROMETERS FOR SIMULTANEOUS DUAL POLARIZATION BIO-CHEMICAL SENSING THROUGH AMPLIFIED FRINGE PACKET SHIFTS ABSTRACT

An optical biosensing set-up is described based on integrated broad-band Young interferometer and capable of projecting on the imaging array two distinct packets of interference fringes, one for TE and the other for TM polarization. With an adlayer growing on the sensing arm, the packet shifts faster than the on-chip phase difference between the two arms providing an amplified optical signal over a range of wavelengths. The set-up incorporates broad-band light sources, external or integrated, the photonic chip with the Young interferometer, and an imaging array. The fringe shifts monitored at the imaging array present a measure of the biomolecular adlayer built-up.

BACKGROUND OF THE INVENTION

Field of the invention

The field of the invention is label-free bioanalytics through interferometric photonic chips. The photonic chips incorporate waveguide based interferometric devices that are personalized through the immobilization of specific recognition molecules. Upon reacting with the analyte molecules, the growing adlayer changes the effective medium sensed by the waveguides causing interferogram shifts. More specifically, instantaneous dual polarization Young interferometry over a broad spectral range is realized on photonic chips and is proposed for the first time as an analytical tool. The invention describes a photonic chip that accepts broad-band light at the input of Young interferometers and projects the interferogram on an imaging array. Interference fringes appear in polarization specific packets that shift with growing adlayer s. On-chip integrated light sources along with slab waveguides provide additional features towards miniaturization and robustens.

Prior art. Miniaturized bioanalytical devices are vital to point of need applications such as personalized health care and detection of harmful substances in food, water and environment. Optical detection provides high sensitivity combined with a wide dynamic range and immunity to parasitic currents and electrical interference. This is a result of the optical frequency regime of operation and the galvanic isolation of the transducer from the excitation and detection electronics. Several optical detection principles, such as surface plasmon resonance (SPR) sensors , optical waveguides, grating couplers, Mach-Zehnder interferometers, microring resonators, resonant mirrors, and reflectometric interference spectroscopy (RIfS) sensors have been explored with regards to the detection of specific analytes at the optical microsystem level. The particular category of planar waveguide based interferometers is far more sensitive compared to free space optics. In an integrated interferometer, the photons in the sensing waveguide probe the biomolecular adlayer hundreds or thousand times compared to a couple of times in reflectometric interference spectroscopy and SPR methods. The sensitivity enhancement and immunity to parasitic effects make planar waveguide based interference devices ideal for label-free testing. As far as the spectral content of the light involved is concerned, traditionally monochromatic light sources are employed. Monochromaticity, though, is not a prerequisite for interferometry. A typical example is the one dimensional case of RIfS [ Petrou, P. S. et al. "Real-time label-free detection of complement activation products in human serum by white light reflectance spectroscopy " Biosens. Bioelectron. 24, 3359-3364 (2009)]. Similarly, broad-band interferometry is possible even in integrated devices aiming at a much better performance than RIfS [Psarouli A. et al " Monolithically integrated broad-band Mach—Zehnder interferometers for highly sensitive label free detection ofbiomolecules through dual polarization optics" Sci. Rep. 5, 17600 (2015)].

Young interferometry has also produced valuable results in refractive index and biosensing measurements. Here, the standard practice is the use of monochromatic light. Few exemptions include the use of two or three discrete wavelengths to enrich the monochromatic device results: [Mulder, H. K. P.; Blum, C. Subramaniam, V., Kanger, J. "Size-selective analyte detection with a Young interferometer sensor using multiple wavelengths" S. Optics Express 24(8) 8594-8619 (2016), Saastamoinen, K., Tervo, J., Turunen, J, Vahimaa, P., Friberg, A. T "Spatial coherence measurement of polychromatic light with modified Young 's interferometer " Optics Express 21(4) 4061-4071 (2013)]. Invention summary and basic advantages

The present invention proposes Young interferometry having as input continuous broad-band light. In a monochromatic Young devices the fringe pattern is a periodic one and has a phase that is the same as the on chip phase difference between the two arms. In the new device, the integration of the interference effects over a wide and continuous spectral range results in structured fringe pattern where interference occurs at specific angles distinct for each polarization. This way two packets of fringes appear on the imaging array, one for the TE and another for the TM polarization. Polarization deconvolution, therefore, becomes straightforward without the use of polarizers. As it turns out, these two fringe patterns move faster than the on chip phase function due to the additional shift of the packet envelope. Additional fringe pattern steering can be achieved by introducing sensing and reference arms with uneven lengths. At the same time, the on-chip phase signal can be measured over a broad spectral range and for both polarizations. The above unique features of the broad-band Young interferometer are all competitive advantages over the monochromatic counterpart. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 : Young interferometer basic configuration and geometrical parameters. After the Y junction, the sensing (1) and reference arms (2) follow. The vertical axis on the right is the imaging array placed at a distance Z from the emitting edges of the two arms. The two arms are separated by a distance D and the sensing window has a length L. Vacuum is assumed between the interferometer and the array.

Figure 2: Young interferometer cross section at the sensing window (a) and at the emitting edge (b,c). The sensing (1) and reference arms (2) are shown as rib waveguides in the sensing window and at the emitting edge in (b). The emitting edge in (c) shows strip waveguides. The waveguide core is the shaded in a,b,c. Also shown is the substrate (3) the top-cladding layer (4) and the bottom-cladding layer (5).

Figure 3. Normalized fringe packets for the TE (right) and TM (left) polarizations. The assumed geometrical parameters were Z=3 cm , Z⁾=200 μπι, L= 1mm

Figure 4. Phase functions, (N-N_s)kL, for TE (lower curve) and TM (upper curve). The average phase function slope for TE is 25X10³ nm and for TM is 89X10³ nm. Here L=\ mm and, therefore, the slope per unit L, becomes a₁=2525X10^"3 and a₂=89X10^~3.

Figure 5. The two polarization packets of Fig.3 after introducing an extra length ^S^C^m on the sensing arm. The rest of the parameters (L,D,Z) remain the same. The packets are now symmetrically placed with respected to the arms.

Figure 6. Fringe packet motion to the right for the TE polarization as a result of a 5 nm adlayer on top of the sensing arm exposed part. The adlayer has a refractive index of 1.45. The broken line is the fringe pattern before the 5 nm adlayer and the solid line the fringe pattern after the adlayer. Here =lmm.

Figure 7. A blow-up near the left edge of the fringe packets for the TE polarization in Fig.6 showing in detail the packet motion to the right as a result of a 5 nm adlayer having a refractive index of 1.45. As in Fig.6, the broken line is the fringe pattern before the 5 nm adlayer and the solid line the fringe pattern after the adlayer. Here again, =lmm.

Figure 8. Fringe packet motion to the right for the TM polarization as a result of a 5 nm adlayer on top of the sensing arm exposed part. The adlayer has a refractive index of 1.45. The broken line is the fringe pattern before the 5 nm adlayer and the solid line the fringe pattern after the adlayer. Here, =lmm.

Figure 9. A blow-up near the left edge of the fringe packets for the TM polarization in Fig.8 showing in detail the packet motion to the right as a result of a 5 nm adlayer having a refractive index of 1.45. As in Fig.8, the broken line is the fringe pattern before the 5 nm adlayer and the solid line the fringe pattern after the adlayer. Here again, Z=lmm.

Figure 10. The absolute value of the phase function change for the TE and the TM polarization as a result of the 5 nm adlayer with a refractive index of 1.45. The adlayer and interferometer optical and geometrical parameters are as in Fig.6. The Coshift, as in Eq.(8), is 11.23 rads and 14.37 rads for TE and TM, respectively. As for the Enshift in Eq.(9), the values are 13.97 rads and 42 rads for TE and TM, respectively.

Figure 11. The two polarization packets for the same waveguide geometry as in Fig.5 but with the interference taking place at the end of an on-chip slab waveguide where the two output waveguides terminate. The length of the slab waveguide is the 3 cm.

Figure.12. Example of a monolithically integrated broad-band light source coupled to a Young interferometer. The broad-band light source is silicon avalanche P⁺⁺N⁺ diode biased beyond its breakdown voltage. The emitting junction is self-aligned to the up-going segment of the silicon nitride waveguide. Shown are the sensing arm (1) the reference arm (2) the Si substrate (3), the top-cladding layer (4) and the bottom-cladding layer (5). The output waveguides terminate at the chip edge. Interferometry occurs off chip. An off-chip lens collimates the beam in the vertical direction and an external imaging array records the spatially distinct TE and TM fringe packets;

Figure.13. Example of a monolithically integrated broad-band light source coupled to a Young interferometer. The broad-band light source is silicon avalanche P⁺⁺N⁺ diode, as in Fig.12. Shown are the sensing arm (1) the reference arm (2) the Si substrate (3), the top-cladding layer (4) and the bottom-cladding layer (5). The output waveguides terminate the on-chip slab waveguide. Interferometry occurs at the chip/slab waveguide edge. An external imaging array in close proximity to the chip edge records the spatially distinct TE and TM fringe packets.

Figure 14. The Young interferometer with the various waveguide structures. The black segments are strip waveguides while the white ones are single mode rib waveguides. The tapers at the strip- rib and rib-strip transition points are required to suppress coupling losses. At the end of the sensing and reference rib waveguides two additional tapers connect to the output strip waveguides. Another taper at the end reduces the strip waveguide width to submicron levels to increase the waveguide numerical aperture. The imaging array (not shown) is supposed to be at a distance Z from the output waveguide edges.

DETAILED DESCRIPTION OF THE INVENTION

Definition of terms

Photonic Chip: A planar stack of dielectric films on appropriate substrates consisting of dielectric bottom-cladding, top-cladding and core layers. The core layer, sandwiched between the bottom- cladding and top-cladding, is patterned so that integrated waveguides form an optical device consisting of the input waveguides, the output waveguides and the interferometer in-between. The bottom-cladding and top-cladding layers are thick enough to isolate the core layer from the substrate and any material in contact with the top-cladding. The waveguide core is made of a higher refractive index material than the bottom-cladding and top-cladding dielectric layers. The core interacts with the environment only in selected areas where the top-cladding is removed. A typical choice would be silicon dioxide for top-cladding and bottom-cladding layers and silicon nitride or silicon oxynitride for core. The substrate is chosen opaque and can be insulating or semiconducting, like silicon, or metal. The photonic chip in certain cases can be upgraded by the on-chip integration of light sources and/or detectors. Light source: A broad-band light source either off-chip or integrated on chip that supplies light of either polarization, TE and TM, into the input waveguides

Input waveguide: The integrated waveguide part that couples broad-band light of both TE and TM polarizations either from an external broad-band light source or from an on-chip integrated broadband light source. Young interferometer: A planar waveguide structure that is a continuation of the input waveguide and where the single waveguide is split into two arms by a Y junction. One arm, sensing arm, has an exposed sensing window having an effective index N_s and a length L. The sensing window of the sensing arm is to be functionalized with the recognition or probe molecule, while the rest of the sensing arm is buried under the top-cladding layer. The other arm, reference arm has an effective index N_r and is usually buried under the top-cladding layer. The light in the two arms experiences different media resulting in a phase difference at the end of the two arms.

Output waveguides: The two integrated waveguide parts that are continuation of the sensing and reference arms following the exposed arm section. The output waveguides radiate the waveguided light at their emitting edges and interference occurs at specific angles depending on the phase function dependence on the wavenumber. The output waveguides are terminated either at the photonic chip edge and radiate to an adjacent imaging array or at an on-chip slab waveguide the edge of which is butt coupled to an imaging array.

Phase function: The phase difference of the fundamental electromagnetic modes travelling through the sensing and reference arms as a function of the wavenumber. Broad-Band Young interferometer: A Young interferometer the input waveguide of which is capable of coupling broad-band light of either polarization. Its reference and sensing arms are properly engineered so that the TE and TM phase functions have distinctively different wavenumber dependence in order for the interference fringe packets for TE and TM polarizations to occur at different and not intersecting angles. Its output waveguides terminate with proper lengths and separation at emitting edges so that the interference fringe packets for the TE and TM polarizations are within the beam divergence angle of the output waveguides.

Array of broad-band Young interferometers: One or more broad-band Young interferometers on the same photonic chip with the all input waveguides on one side and all the output waveguides on another side. Light source: A broad-band light source either off-chip or integrated on chip that supplies light of either polarization, TE and TM, into the input waveguides.

Imaging Array: A digital monitor of the total light spatial intensity in the form of a CCD array or a CMOS pixel array.

Sensing window: The exposed part of the sensing arm in contact with the sample or reagents and functionalized by immobilizing the probe molecules on it.

Probe molecules: The molecules that are coated on the sensing window. They can be proteins, hormons, DNA fragments, or other types of molecules that specifically react with a counterpart molecule of interest.

Photonic Biochip. A photonic chip with one or more interferometers and sensing windows where one or more recognition molecules have been immobilized.

Analyte molecule: A molecule of analytical interest that will react specifically with probe molecule immobilized on one of the sensing windows of the photonic biochip.

Basic Equations and Concepts

To derive an approximate analytical expression for the broad-band Young interferometer, we start with the standard equation for the classical monochromatic Young interferometer shown in Fig. 1, as a top view and in Fig.2 as a cross section. The photon flux recorded at each point x of an imaging array located at a distance Z from the waveguide emitting edges which are separated by a distance D from each other is given by the following equation:

where //, h are the photon fluxes from the sensing and the reference arms, respectively, k is the vacuum wavenumber, k=2n/ , and φ(Κ) is the phase difference between the two arms. This phase difference is as a function of the wavenumber and will be called phase function. In the above equation the cosine term implies that the monochromatic interferogram has a period of 2nZ/(Dk).

In the case of broad-band Young Interferometry, a wide range of photons with wavenumbers from ki to k2 interfere at the same time. Wavenumbers ki and correspond to the upper and lower wavelengths, λ₂ and λι, of the waveguided spectrum, respectively. The total photon flux sensed at a location x on the screen due to all wavelengths of the spectral region is the integral of the various photon fluxes. Assuming the fluxes to be constant with the wavenumber between ki and we obtain for the total photon flux, S(x)

xDk

S(x) = (k₂ - k_x )(/, + 1₂ ) + lfi ₂ ^~ j cos(^— + (p(k))dk (Eq. 2)

To integrate the last expression in closed form and for the sake of simplicity we will assume the phase function <p(k) to be linear with the wavenumber k:

<p(k) = [(N_r -N_s)L]k = (φ_0ί +a_tk)L (Eq. 3) where N_r and N_s are the reference and sensing arm effective indices, respectively, and L is the sensing window length. The linearity hypothesis will be later relaxed. In Eq. (3) a and (pot is the slope and the zero k intercept of the phase function per unit L. Subscript i=l,2 stands for polarization, 1 for TE and 2 for TM. Given the linearity assumption the S(x) integral can be derived in analytical terms as Sl(x) given by

Sl(x) = {k₂—k ) I_x +I₂ + 2 VV^", O) cos[⁽* ^{+ g}' ^^{/ )) )} (k₂ + k,) + <p_0lL] (Eq. 4) where the cosine term has a period equal to 4nZ/[D(ki+k2)] and E (x) is the envelope function of the cosine term in the previous equation:

(x + d_jLZ / D)D

sin[-

En^x) = 2Z (½ -*,)]

(Eq.5)

(χ + α,ΣΖ / D)D

(k₂ k_x )

2Z

Polarization Packets

By comparing the first expression for Ι(λ) (monochromatic Young interferometer) with the last two equations for Sl(x) (broad-band Young interferometer) one realizes that the cosine terms are quite similar with the wavenumber k and the phase difference term φ being replaced by average values for the wavenumber, (ki+k2)/2, and also for the phase: (<poi+ai(ki+k2)/2)L. What is strikingly different, though, is the envelope function, E_n(x) as a result of the broad-band nature of the interference pattern.

The envelope function is of the sinx/x type and centered around -aiLZ/D. Therefore, each polarization appears as a packet centered at an angle -aiL/D with respect to the symmetry axis between the two waveguides at the emitting edge. It is assumed that the absolute value of aiL/D is much less than 1. Here it is mentioned that aiL is the slope of the phase function plotted against k. The main lobe of the Eni(x) function extends ±2nZ/[(k2-ki)D] around maximum, or between angles ±2n/[(k2-ki)O] with respect to the angle of maximum, assuming 27i/[(A:2-A:/)D]«l . With proper selection of the of the slopes ,, and parameters L and D, we can make the separation of the two polarizations packets the equivalent of several side lobes so that the two packets are clearly split apart and separated from each other.

Packet Steering

In addition to the polarization splitting , the two interference fringe packets can be steered so that they are placed symmetrically from either side of the origin (x=0), while keeping their separation. This way they lie well within the beam divergence angle as defined by the waveguide numerical aperture in the horizontal plane. This is done by introducing uneven lengths in the reference and sensing waveguides away from the sensing area. By choosing the reference waveguide length to be shorter than the sensing one by SS, while keeping the same exposed arm length L, the phase function becomes

φ{Κ) = [(N_r - N_s )L]k - SSN_rk = (φ_0ί + a_tk)L - SSN_rk (Eq. 6)

Provided that N_r is a slowly varying function of k, the phase function assumes another slope:

SSN_r instead of a L. Consequently, packet steering is made possible by a distance of SSN_rZ/D and the new polarization packets are centered at angles -aiL/D+ SSN_r ID. The magnitude of these angles can be made smaller than the output waveguide numerical aperture NA_owg by properly choosing, 5S, L and D, so that the two packets are emitted within the waveguide beam divergence angle. In more analytical terms SS must be chosen so that :

For a given waveguide geometry, that defines a, and NA_owg, and for a given exposed arm length L, that defines sensitivity, SS and D can be chosen to fulfill two criteria: First, to satisfy the above inequality, and second, to keep the interferogram within a readable range by ensuring that its period, 4nZ/[Dflti+k2), is well above the imaging device resolution.

Phase Function and Fringe Shift Sensitivities

The effect of the adlayer growth on the envelope function and fringe shift comes from its influence on the slope of the phase function, α,Ζ. An adlayer of thickness d forming on the sensing arm changes the phase function by S<p(k). By assuming, again, linearity

δφ(Κ) = -d(dN_s I dd)Lk = -d(<p_0ld + a_idk)L (Eq. 7) where φοΐά and aid express the sensitivity of the sensing arm effective index to adlayer growth. By replacing in the expression for Sl(x),Eq.(4), the new phase function we obtain a total fringe pattern shift consisting of two terms: The cosine shift, Coshift, and the envelope shift, Enshift. Expressed in angle terms, where a displacement by one fringe equals a 2π, these terms take the form:

Coshift=- d[q>_0id + a_id (k₂ +k l 2]L (Eq. 8)

Enshift- da_id (k₂ + k, )L I 2 (E - ⁹) The first quantity is identical as the average of the on chip phase change, δφ, due to the adlayer. The minus sign in either term comes from the fact that the absolute value of the phase function decreases with the increasing effective index of the exposed sensing arm. It turns out that the envelope shift can be several times the cosine shift because the terms φοίά and m(ki+k2) have different signs while comparable in magnitude.

By applying Fourier transform to the function on the right hand side of the expression for Sl(x), Eq. (4), we obtain a finite and constant magnitude from kiD/Z to foD/Z while zero everywhere else. In the Fourier transform domain the variable is k'=&D/Z and the phase changes linearly from (φ ί+aiki )L at

The average value of the phase shift due to the adlayer as calculated through Fourier transform is the Coshift term, Eq. (8). An apparent total shift is obtained if the envelop shift, Enshift, is added. This shift can be calculated after the envelope function is deconvoluted from the recorded patterns through signal processing algorithms, including frequency shift techniques. Therefore, by summing the Coshift and Enshift an "amplification" factor, SA, relative to δφ is obtained:

Coshift + Enshift

SA = ¾ (*j ⁺ *2 ) + (Eq. 10)

Coshift EXAMPLES

Two Distinct Polarization Packets

To test the validity of the two distinct polarization packets in actual devices and after relaxing the phase linearity hypothesis, we employed numerical simulations for the phase function in devices with realistic geometrical and optical parameters. The phase function (p(k) was numerically calculated as (N_r-N_s)kL. The fundamental mode effective indices N_r, N_s were obtained through mode simulations with the FemSIM software package (SYNOPSYS) for the employed waveguide geometries and the standard material refractive indices for the nitride core, oxide claddings and assay buffers. Here, we made use only of the basic interference equation and numerically calculate the integral of S(x) as in Eq. (2). The integral extended from 650 nm (2π/&₂) to 850 nm The numerically calculated results exhibit patterns with distinct polarization packets as shown in Fig. 3. The assumed geometrical parameters were Z-3 cm, D=200 μτη, L= lmm. The reference arm has silicon dioxide top and bottom-cladding layers while the exposed sensing arm has as a top-cladding layer the assay buffer and silicon dioxide as lower cladding. The waveguides were silicon nitride monomodal rib waveguides with a width of 1 μηι, an etch depth of 6 nm and a thickness of 135 nm for the sensing arm and 150 nm for the rest of the waveguides including the reference waveguide. The difference in the two thicknesses is desirable for the phase functions to have distinctly different slopes ai and c¾ (Fig. 4) for a clear polarization packet splitting. The distance between the two packets is due to the slope factor a that is much higher for TM (<¾=89X10^"3) than for TE (ay =25X10^"3) as shown in Fig.4. The packet center position obeys the -aiLZ/D law. For =1000 μιη, Ζ=30000μηι (3 cm) and D= 200 μηι, the two positions are 3750 μηι (TE) and 13300 μηι (TM), in agreement with Fig. 3. The fringe oscillation period is 112 μιη, in agreement with the term 4 Z/[D(ki+k2)] derived from the expression for Sl(x), Eq. (4). As for the envelope width, the numerically simulated packets are wider than the twice the value of 2nZ/[(k2-ki)D], obtained by the assumption of the linear phase function. This is a result of the deviations of the phase functions from the strict linearity condition. The particular choice for the waveguides thickness (135 nm sensing, 150 nm reference arm) was just an example and is not unique to obtain polarization packet splitting. Different thinning degrees can also work, as well as thinner waveguides for each arm.

Packet Steering to Symmetric Position

The middle point between the two polarization packets in Fig. 3 was about 8500 μηι left of the x=0 point. Based on the packet steering discussion, an introduction of a reference waveguide that is shorter by <55^,=8500(D/Z)(l/N_r) μπι compared to sensing waveguide should bring the two packets symmetrically across the x=0 point. The length SS is calculated at about 30 μηι using the parameters presented in the previous section and indeed brings the two packets symmetrically across the x=0 point, as verified in Fig. 5. The introduction of uneven arms in no way implies that the straight line segments of the sensing and reference waveguides at the exposed sensing arm area will be redesigned. Away from the exposed sensing arm , an extra length of the order of few tens of a micrometer can be introduced at the sensing waveguide in various ways, for example by increasing the arm's curvature radius at the Y junction.

Accelerated Fringe Packet Shifts

To demonstrate the accelerated fringe packet shift as a result of adlayers we calculated the phase function change, δφ ), as well as the actual fringe packet shift upon adlayer growth and for the same waveguide parameters and interferometer geometries as in Fig. 5. Again the reference waveguide was 30 μπι shorter than the sensing one, to center the fringe packets. The adlayer was a biomolecular film with an optical thickness of 5 nm and a refractive index of 1.45. The results are shown in Figs. 6-9. The entire TE and TM packets is shown in Fig.6 and Fig.8, respectively, while Fig.7 and Fig.9 show in more detail the packet edge shifts for the TE and TM, respectively. According to Eq.8 and Eq.9 the adlayer shifts the two packages to the right by a number of periods equal to Enshift/27r. Considering the slopes of the adlayer induced phase function changes from Fig. 10, then through Eq.8 and Eq.9, Enshift^ is calculated at 2.22 periods for TE and at 6.68 periods for TM. These values agree with the observed adlayer induced shifts in Figs. 6-9. This is to be contrasted with the fringe shift Coshift^ values of 1.79 periods for TE and 2.28 periods for TM. The relevant amplification factors, SA, are 2.2 and 3.9 for TE and TM, respectively.

The broad-band Young interferometer concept presented so far also apply to chips with on-chip interference across the free propagating region (slab waveguide) instead of vacuum. Here, the output waveguides terminate at the onset of the on-chip slab waveguide which serves as the interference medium while its other end is butt-coupled to an imaging array. In the expression for S(x), Eq.(2), the vacuum wavenumber in the cosine term is now multiplied by the effective index of the slab waveguide. Very similar, though narrower, patterns are produced and the two polarization packets are still separated with smaller packet angles. Such interference fringes and associated polarization packets are shown in Fig. 11 with the same arm and waveguide geometry, separation and array distance as in Fig.5.

Monolithically Integrated LEDs and interference in air.

An example of a broad-band Young interferometer coupled to a monolithically integrated broadband light source is given in Fig. 12. Here, the output waveguides terminate at the chip edge. The light source is a P^"1^^4" avalanche diode which emits broad-band light when biased beyond its breakdown voltage. The up going segment of the silicon nitride waveguide serves as the input waveguide and is self-aligned to the diode metallurgical junction so that high coupling efficiencies are obtained between the input waveguide and the avalanche diode LED. [Psarouli A. et al "Monolithically integrated broad-band Mach—Zehnder interferometers for highly sensitive label free detection ofbiomolecules through dual polarization optics" Sci. Rep. 5, J 7600 (2015)]. Both the TE and TM polarizations are excited at the input waveguide. The light emitted by the output waveguide emitting edges, after being collimated in the vertical direction by a lens, is reaching an imaging array where the two beams (sensing and reference) interfere and form two distinct polarization packets, one for the TE and another for the TM polarization.

Monolithically Integrated LEDs and slab waveguide interference.

An example of a broad-band Young interferometer with on-chip interference is given in Fig. 13. A monolithically integrated broad-band light source is coupled to a Young interferometer the output waveguides of which terminate at the on-chip slab waveguide. The light source is the P⁺⁺N⁺ avalanche diode of the previous section which emits broad-band light and is self-aligned to the input waveguide. Both the TE and TM polarizations are excited at the input waveguide. Interference occurs at the slab waveguide-chip edge and is captured by an imaging array that it butt-coupled at close proximity to the chip edge,

Broad-band Young interferometer layout

A more elaborate layout of the Young interferometer waveguides is presented in Fig. 14. At the beginning the waveguide starts as a fully etched strip waveguide, as dictated by the need of self- alignment with the avalanche diode metallurgical junction in the case of monolithically integrated LEDs. It then connects to the rib waveguide by a narrow- wide-narrow taper (Fig. 14) to reduce coupling losses. The rib waveguide section includes the Y-j unction and the reference and sensing arms. It is shallow etched to ensure monomodality and low propagation losses. The optimum Si₃N₄ core thickness is chosen using the criteria of high sensitivity and clear separation of the two polarization packets. Throughout the rib waveguide the two arms distance is kept at a range of few tens of a micron to limit S bent lengths. Following the sensing and reference arms, the output waveguides follow as strip waveguides after another narrow-wide-narrow taper intervenes. Given the concepts developed in the previous sections, the separation of the two output waveguides at the emitting edges must be wide enough for the two polarization packets to appear within the waveguide beam divergence angle. Separation values in the hundreds of a μηι are anticipated. Such a wide separation of the two output waveguides is accomplished at a relatively short distance since the strip waveguide geometry allows for much steeper bents. Also, a standard taper at the strip end reduces the strip waveguide width to submicron levels which increases the waveguide numerical aperture and re-introduces monomodality. The increased numerical aperture is required to ensure that both polarization packets are within the waveguide beam divergence angle.

ADVANTAGES Monochromatic Young interferometers have so far provided the lowest detection limits of effective index changes either due to cover medium changes or due to adlayer formation as a result of biomolecular binding. The basic configuration is simple: Through a Y junction a waveguide is split into two arms, a reference and a sensing one. The two arms are terminated at the chip edge and emit the waveguided light towards an imaging array. Due to interference of the two diverging beams, the light intensity on the screen varies sinusoidally with a period that is proportional to the wavelength times and the distance from the screen and inversely proportional to the separation between the two arms. The sinusoidal fringe pattern has a phase that is the same as the phase difference of the light propagating into the two arms. Thus any change in the effective index of the sensing arm will be recorded as a change into the sinusoid phase. Given the excellent imaging devices available, CCD or CMOS pixel arrays, and the ability to scale the signal period through the interferometer-imaging array geometrical factors very low noise levels are achievable. At the same time phase signal sensitivities increase linearly with the exposed arm length so that low limits of detection are achievable, in fact the lowest among similar interferometric devices. Cover medium RI limits of detection below 0.5X10^"7 RIU are achievable with such interferometers. Given the inherent advantages of the standard monochromatic Young interferometers, the present invention envisages to improve even further the device performance by coupling into the input waveguide broad-band light of both TE and TM polarization. Such input light introduces another dimension into the observables and this is the instantaneous and independent measurement of the phase signal for the two polarization over a range of wavelengths. This way the effects of cover medium changes can be deconvoluted from the adlayer and the adlayer refractive index can be obtained without using any external polarizers and additional measurementas as in monochromatic Young interferometry. This is possible as the dispersion relationship of the phase function with the wavenumber is substantially different for the two polarizations and this, in turn, results in two spatially distinct fringe patters at the imaging array site. The use of broad-band light enables a wider variety of light sources to be employed compared to the laser sources used in the standard monochromatic Young interferometer. As an additional feature, the fringe patterns are enclosed in envelope functions that move with adlayer growth. In fact, the envelopes move faster than the sinusoidal signal and act as an optical signal amplifier.

Claims

1. A photonic chip comprising:

an array of broad-band Young interferometers;

reference and sensing arms for each interferometer properly engineered so that the TE and TM phase functions have distinctively different wavenumber dependences in order for the interference fringe packets for TE and TM to occur at different and not intersecting angles;

output waveguides terminated at the chip edge and with a proper lengths and separation at emitting edges so that the interference fringe packets for the TE and TM polarizations are within the beam divergence angle of the output waveguides.

2. The photonic chip of claim 1 where to each input waveguide a broad-band light source is coupled from an on-chip integrated array of broad-band light sources.

3. A photonic chip comprising:

an array of broad-band Young interferometers;

an on-chip array of slab waveguides of appropriate dimensions, each for every interferometer; output waveguides terminated at the slab waveguides onset and with proper lengths and separation at emitting edges so that the interference fringe packets for the TE and TM polarizations are within the beam divergence angle of the output waveguides and a readable interferogram forms at each of the slab/chip edges.

4. The photonic chip of claim 3 where to each input waveguide a broad-band light source is coupled from an on-chip integrated array of broad-band light sources.

5. An apparatus for the detection of analytes consisting of:

the photonic chip of claim 1 ;

an array of off-chip broad-band light sources each coupled to one of the input waveguides;

a single imaging array placed against all the emitting edges of the Young interferometers output waveguides and at a proper distance to monitor readable interferograms showing the distinct TE and TM fringe packets;

probe molecules immobilized on the sensing windows of the Young interferometers to create fringe and packet shifts on the interferogram upon binding with the analyte molecules by immersing the sensing windows into the analyte solution;

bias and control electronics for the operation of the off-chip broad-band light sources and the imaging array and for the multiplexing of the light sources so that all integrated interferometers are independently interrogated.

6. An apparatus for the detection of analytes consisting of:

the photonic chip of claim 2;

bias and control electronics for the operation of the integrated array of broad-band light sources and the imaging array and for the multiplexing of the light sources so that all interferometers are independently interrogated.

7. An apparatus for the detection of analytes consisting of:

the photonic chip of claim 3;

a single imaging array placed in close proximity to the slab waveguides/chip edge to monitor the interferograms showing all the interferometers distinct TE and TM fringe packets;

bias and control electronics for the operation of the off-chip light sources and the imaging array and for the multiplexing of the light sources so that all integrated interferometers are independently interrogated.

8. An apparatus for the detection of analytes consisting of:

the photonic chip of claim 4;

probe molecules immobilized on the sensing windows of the Young interferometers to create fringe and packet shifts on the interferograms upon binding with the analyte molecules by immersing the sensing windows into the analyte solution;

bias and control electronics for the operation of the integrated array of light sources and the imaging array and for the multiplexing of the light sources so that all integrated interferometers are independently interrogated.