1. Introduction

Chemical and biological sensors based on integrated optical (IO) Young interferometry have attracted considerable attention in the past decade because of their high sensitivity, label-free detection capability and large dynamic range [1–11]. Several types of IO Young interferometer sensors have been developed based on either slab or stripe waveguides. There are two slab Young interferometer sensors, one consists of two planar core layers separated by a low-index buffer layer, which is now commercially available [1–4]; the other uses double slits to yield two light beams that were simultaneously coupled with an integrated grating into a single planar waveguide for sensing and reference [5,6]. The stripe Young interferometer sensors generally contain a sensing and a reference arm that are connected with a 3dB Y-branch power splitter (an exception is shown in Refs. 9, 10) [7–10]. For these existing IO Young devices, the butt-coupling method was commonly used and the measurand-induced phase variation was generally determined by means of Fourier analysis of the spatial interference pattern detected with a linear charge-coupled device (CCD) detector.

We recently developed a simple IO Young interferometer sensor that is different from the conventional devices in four aspects [11]: (1) the sensor chip is a glass sheet with several pairs of single-mode straight channel waveguides but without having a Y-branch power splitter, such a simple structure makes the sensor chip readily prepared for a single use; (2) the sensing arm is a channel-planar composite waveguide that offers the sensor a high sensitivity and enables the uncovered channel waveguide to be used as the reference arm; (3) the prism-coupling method was used, to provide the sensor with a high-contrast dotted-line interference pattern in the lateral direction; (4) by using a simple single-slit photodetector instead of the expensive CCD detector, the induced variation in the phase difference of the sensor was detected in real time.

However, we found in the previous work that the separate components such as the prism couplers and the fluid chamber make the construction and operation of the sensor system rather cumbersome, despite of the integrated waveguide chip. To overcome this drawback, a novel prism-chamber assembly was prepared. The prism-chamber assembly facilitates the replacement of the waveguide chip, and provides the position-limited tight contacts of both the prism couplers and the fluid chamber with the waveguide chip, leading to an easy and reproducible operation of the sensor device. By using the prism-chamber assembly, the performance of the IO Young interferometer sensor was systematically investigated. Since the prism coupler and the fluid chamber are commonly used in a large variety of waveguide-based chemical and biological sensors, the prism-chamber assembly proposed in the present work is also be applicable to other waveguide sensors for a convenient use. The experimental data obtained with the prism-chamber assembly were shown below.

2. The waveguide chip and the prism-chamber assembly

The structure and fabrication of the waveguide chip were described in Ref. 11. Figure 1(a) shows a photograph of the sensor chip in which 5 pairs of single-mode straight channel waveguides were prepared by ion exchange in glass. Each channel pair can be used as an IO Young interferometer, hence a single chip contains an interferometer array. All channels are 4 μm wide and 45 mm long. The distance between the two neighboring channel pairs is 1 mm, much larger than the channel-to-channel spacing of d = 75 μm within a pair [see Fig. 1(b)]. The bright lines in Fig. 1(a) are the sensing arms of the interferometer array, which were fabricated by selective-area sputtering of the tapered layer of TiO₂ on the chip. The uncovered channels serve as the reference. The bright spot in Fig. 1(c) is the near-field modal profile observed using a fiber to couple the 532 nm laser light into a single channel. Figure 1(d) shows the far-field interference pattern observed after simultaneously launching a 633 nm laser beam into a pair of channels with the prism coupling method. The dotted-line pattern was formed due to overlapping of the laterally divergent light beams coupled out of the channels via another prism. The high contrast of the pattern indicates the equal intensities of the two output light beams, being ascribed to the flexible prism-coupling method.

 

Fig. 1 (a) Photograph of the glass chip with 5 pairs of single-mode channel waveguides; (b) Optical microscope image of the channel waveguide pair; (c) Near-field modal profile at λ = 532 nm of a single channel; (d) the far-field interference profile produced from the channel waveguide pair with the prism-coupling method; (e) Young interference pattern simulated with the parameters of the actual sensor (the blue curve is the diffracted light intensity profile for a single channel).

Download Full Size | PDF

To facilitate operation of the IO Young interferometer sensor and to improve the sensor performance, the prism-chamber assembly was designed and prepared. Figure 2(a) shows the front view of the prism-chamber assembly in the close state; Fig. 2(b) illustrates the side view of the assembly in the open state. Figure 2(c) shows a photopicture of a real prism-chamber assembly being used. The prism-chamber assembly consists of an aluminum back board for holding the waveguide chip, a PMMA front board in which two fluid chambers with the respective inlets and outlets were made (each chamber 300 μL), the silicone gasket, the prism in- and out-couplers that were mounted on their holders. The prism holders were flexibly linked to the front board with the slide screws. The back and front boards are interconnected at their bottom sides with the axis of rotation. When using the prism-chamber assembly, just place the waveguide chip on the given area of the back board and then tightly sandwich the chip between the back and front boards with the locking screw. The two spring between each prism holder and the front board lead to a constant-pressure attachment of the prism coupler to the waveguide chip. With a peristaltic pump the sample solution is injected in the chamber at a given rate. The silicone gasket offers the chamber a good sealing. A miniaturized thermal sensor was installed in the chamber to detect the solution temperature. A linearly polarized laser beam of 633 nm wavelength was irradiated upon the prism in-coupler at a given angle, to simultaneously excite the fundamental transverse electric (TE) or magnetic (TM) modes in a pair of channel waveguides. A single-slit photodetector was fixed at a distance of D = 95 mm from the prism out-coupler, and the sensor system was investigated. With λ = 633 nm, D = 95 mm and d = 75 μm the spatial interference pattern of the sensor was simulated, as shown in Fig. 1(e). The spatial periodicity was calculated as λD/d = 0.8 mm.

 

Fig. 2 (a) the front view of the prism-chamber assembly in the close state; (b) the side view in the open state; (c) Photograph of an actual prism-chamber assembly (1. back board, 2. waveguide chip, 3. prism coupler, 4. prism holder, 5. springs, 6. front board, 7. slide screws, 8. fluid chamber, 9. inlet and outlet, 10. locking screw, 11. strut, 12, rotation axis, 13. silicone gasket, 14. temperature sensor)

Download Full Size | PDF

3. Influence of the slit width on the fringe contrast of the sensor

Fringe contrast is an important parameter with the interferometric sensor. The higher the fringe contrast, the larger the resolution in the phase-difference change, and the lower the detection limit. Thus, a large fringe contrast is always desirable for the interferometric sensor. In this work, the fringe contrast of the IO Young interferometer was investigated versus the slit width of the slit photodetector. To change the phase difference with time, the liquid in the chamber was slowly exchanged between water and the aqueous NaCl solution (11 wt%). Figures 3(a) –3(d) show four temporal interference patterns corresponding to the slit widths of a = 0.2 mm, 0.5 mm, 1.0 mm, and 1.5 mm, respectively. With the a = 0.2 mm and 0.5 mm slits the regular interference patterns with the fringe contrasts of γ = 0.76 and 0.37 were obtained. Increasing the slit width to a = 1.0 mm and 1.5 mm leads to the deformed interference patterns with the fringe contrasts as low as γ ≈0.11 and 0.05. For comparison, the fringe contrast versus the slit width was calculated by substituting λ = 633nm, D = 95 mm, d = 75 μm in Eq. (1) that was theoretically deduced in the previous work [11]:

 

Fig. 3 (a)-(d) Temporal interference patterns of the sensor measured with the prism-chamber assembly and the slit detector of different slit widths [(a) 0.2 mm, (b) 0.5 mm, (c) 1 mm, (d) 1.5 mm], (e) the fringe contrast versus the slit width (blue curve: the calculated data; red squares: the experimental data).

Download Full Size | PDF

As shown in Fig. 3(e), the fringe contrast decreases down to γ = 0 with increasing the slit width up to a = 0.8 mm that is equal to the spatial periodicity. When the slit width is beyond the spatial periodicity, the fringe contrast becomes small (γ < 0.22). The squares in Fig. 3(e) represent the experimental data, which are little smaller than the calculated values at each slit width. The findings reveal that the fringe contrast of the IO Young interferometer sensor is closely related to the spatial periodicity. By using a simple slit photodetector with the slit width being smaller than the spatial periodicity of the sensor, the regular-shape and high-contrast temporal interference pattern can be easily obtained.

4. Response of the sensor to temperature

For chemical and biological sensing application, the IO Young interferometer is required to be insensitive to the ambient temperature around the device. Here the response of the sensor to temperature of water in the chamber was investigated. To avoid an abrupt refractive-index change, the water was pumped into the chamber prior to heating. The water temperature in the chamber was monitored in real time by mounting a small thermo-couple probe in the chamber. In the meantime, the induced phase-difference change was detected with the = 0.2 mm slit. Figure 4(a) shows a reversible response to water temperature. In Fig. 4(a) the water temperature increases from 28.6°C to 50°C, leading to a phase-difference change of Δϕ = −194.7°. The thermo-optical coefficient of water is −0.000091/°C, thus a temperature change of ΔT = 21.4°C would lead to a refractive-index variation of Δn = −0.00195, which corresponds to a phase-difference change of Δϕ^/ = −131.5° according to the refractive-index sensitivity measured with the same waveguide chip. The difference between Δϕ and Δϕ^/ is −63.2°, which results from the difference in the thermo-optical effect between the sensing and reference arms of the chip. Figure 4(b) shows Δϕ, Δϕ^/ and Δϕ − Δϕ^/ versus T. With the quasi-linear dependences of Δϕ^/ and Δϕ − Δϕ^/ on T the slopes were determined as d(Δϕ^/)/dT = −6.14°/°C and d(Δϕ−Δϕ^/)/dT = −3.26°/°C. The phase-difference changes contributed from the thermo-optical effect of water and that of the chip have the same signs but different values. The former is almost 2 times as large as the latter. Owing to the same signs of the two contributions, the temperature influence on the sensor’s sensitivities cannot be ignored. For accurate measurement, the temperature of the solution sample should be controlled to be equal to the ambient temperature. If the measurement needs to be performed at an elevated temperature, then the temperature influence on the measured result should be removed using the T−Δϕ calibration curve shown in Fig. 4(b).

 

Fig. 4 (a) Time variation of the water temperature and the response of the IO Young interferometer to water temperature; (b) Δϕ, Δϕ^/ and Δϕ − Δϕ^/ versus water temperature (Δϕ is the measured phase-difference change, Δϕ^/ is that calculated due to the thermo-optical effect of water.)

Download Full Size | PDF

Figure 5(a) shows response of the sensor to β-casein adsorption from the aqueous protein solution (3 μM). It is seen from this result that a small change of Δϕ = 2° per minute can be detected. A combination of Fig. 4(a) and Fig. 5(a) indicates that with the prism-chamber assembly the sensor is stable and reliable, capable of real-time detection of very slow physical and chemical processes at the waveguide surface.

 

Fig. 5 (a) Response to β-casein adsorption of the IO Young interferometer with the TE mode (b) Response to refractive index of liquid of the same sensor with the TM mode.

Download Full Size | PDF

5. Polarization dependence of sensitivity of the sensor

The above measurements were fulfilled with the TE mode. Compared with the TE mode, the TM mode makes the same waveguide chip much less sensitive to refractive index of liquid and thickness of the adlayer. This is so because a small thickness of the tapered layer of TiO₂ on the sensing arm makes the evanescent field with the TM mode much weaker than that with the TE mode. Figure 5(b) shows the sensor response to liquid index with the TM mode. Δϕ induced by the water−salt solution exchange is smaller than 180°. In contrast, with the TE mode Δϕ was measured to be 24 × 180° [11]. Therefore, the refractive-index sensitivity with the TE mode is 24 times greater than that with the TM mode. In this work, the thickness of the tapered film of TiO₂ was controlled to be below 30 nm. Simulation based on a five-layer planar waveguide model indicates that for a high sensitivity with the TM mode, the thickness of the tapered layer of TiO₂ should be increased in the range of 40 to 60 nm. In this case, the sensor could not work with the TE mode because of a tremendous scattering loss near the cut-off position in the first TiO₂ taper [12].

6. Conclusions

The IO Young interferometer sensor with a prism-chamber assembly and a single-slit photodetector was characterized. The fringe contrast depends on the ratio of the slit width to the spatial periodicity of the sensor, the regular high-contrast interference pattern can be readily detected as the slit width is smaller than the spatial periodicity. The thermo-optical effect of water together with that of the sensor chip makes the temperature influence on the sensor’s sensitivity unnegligible. Depending on the thickness of the tapered layer of TiO₂, the sensitivity with the TE mode is different from that with the TM mode. The prism-chamber assembly offers the sensor a good stability, and the sensor allows for continuous measurement over several hours.