Open Access
Add to BibSonomyAbstract
Various functional optical devices are integrated on a single chip in order to construct optical current transducers based on polarization rotated reflection interferometry, which consists of polarization maintaining 3-dB couplers, TE-pass polarizers, TE/TM polarization converters, and thermo-optic phase modulators. By virtue of the device integration, the sensor exhibited good linearity, and excellent accuracy with an error less than 0.2%. The integrated-optic device provides inherent polarization maintaining characteristics and precise controllability of the optical path length in the interferometric sensor. Single chip integration reduces the complexity of the interferometry, and enables mass-production of low-cost high performance current sensors.
©2011 Optical Society of America
1. Introduction
The sources of electric power have diversified from the traditional fossil fuels to renewable energy sources such as wind, solar, and hydro energy in order to reduce pollution and slow down global warming. However, the renewable power sources could not generate electric power continuously, and various power sources need to be connected to form a power grid network so as to maintain an unceasing supply of electric power to meet consumer demands. In such a power grid network, the real-time monitoring of electric current at each node is necessary for the maintenance of the network as well as for the prevention of accidental overflow into a certain node before a catastrophic power failure occurs over a wide area.
Optical current sensors have been widely used [1–3] for monitoring the large current in power grid networks. Compared to electric current transducers, optical current sensors have many advantages such as higher accuracy for power metering applications, a wide dynamic range for current measurement, and a large signal sensing bandwidth when used for monitoring the transient response due to a surge current. Since the optical sensors are made up of dielectric materials and have no oil or gas, HV isolation requires less effort. Lightweight optical sensors can be installed easily and require less maintenance [4,5].
The early version of the optical current transducers (OCT) had a difficulty to provide sufficient reliability when it was operated in harsh environments. To overcome the stability issue, the adoption of polarization rotated reflection interferometry (PRRI) has been required, while it has increased the number of optical devices for OCT [1]. To resolve the complexity of the PRRI-OCT, we propose integrated optics to play major role in order to facilitate the production of PRRI based OCT. The building blocks of the OCT as polymeric waveguide devices have already been demonstrated [6–8]. We have recently demonstrated the PRRI-OCT by assembling several polymeric waveguide components along with fiber-optic components in order to assure that the polymer devices are appropriate for constructing the reflection interferometry [9]. The integrated sensor with low-birefringence optical waveguide was also reported [10].
In this research, we have accomplished the integration of various optical components required for PRRI on a single integrated optic current transducer (IOCT) chip in terms of the polymeric waveguide components. The integrated-optics enables large volume production of optical devices as the electronic integrated circuits so as to achieve significant reduction of production cost. The uniformity of device characteristics is much superior to that of the fiber-optic OCTs because the integrated devices have superior uniformity in the production, and the polarization is manageable between the devices. In the integrated device, however, the individual device characteristics could not be measured independently because the devices are connected to each other. Hence, the final resultant output signal has to be investigated as a function of the input polarization states in order to confirm that all the combined devices are functioning properly.
2. Operating Principle of IOCTs
In the current sensor based on the PRRI, to measure the phase difference imposed on the two orthogonal polarized modes, devices for manipulating the polarization states of the guided modes are necessary along with optical splitters and optical phase modulators. The configuration of the PRRI is drawn in Fig. 1 , which includes the following: an IOCT chip with two directional-couplers (DCs), two TE-pass polarizers, a thermo-optic phase modulator, a half-wave plate (HWP), and waveguides connecting the functional components. Superluminescent laser diode (SLD) with a broad spectrum is connected at the input port of the IOCT to excite 45° linearly polarized light so as to excite both TE and TM polarizations. At the output of IOCT, a PM fiber is connected to deliver the optical signal to the fiber sensing coil. In the coil, the Faraday effect is occurred due to the magnetic field induced by electric current flowing through the transmission lines under measurement.
Fig. 1 Configuration of the current sensors incorporating the integrated optical current transducer chip. The chip consists of (1) two directional couplers (DC1 and DC2), (2) two waveguide polarizers (Pol), (3) a half-wave plate (HWP), and (4) two phase modulators (PM1 and PM2). The process of polarization conversion at each step is denoted in the figure.
The input light from SLD that is coupled to the TE and TM modes of the waveguide passes through the IOCT and excites the electric fields parallel to the fast and slow axes of the PM fiber. These field components are changed to circular polarization by the quarter-wave plate (QWP) located in front of the fiber sensing coil. The circular polarizations are the eigenmodes of the sensing fiber under the influence of the magnetic field. Therefore, when the circularly polarized field components propagate through the coil, there is a phase shift between them, but no coupling. A mirror attached at the end of the coil reflects the circularly polarized waves and exchanges their polarization states, i.e., RHCP to LHCP and vice versa. When the reflected light waves propagate back to the input, the phase shift doubles because the direction of the magnetic field changes with respect to the propagation direction.
When the reflected light passes through the QWP again, RHCP converts into a fast axis component and LHCP to a slow axis component. Therefore, the initial polarization components are exchanged when they are returning to the IOCT, i.e., the initial TE component is changed to TM component on its return, and vice versa. Consequently, the inherent phase difference between the two polarizations caused by the birefringence of the PM fiber is canceled out. In a similar manner, any additional birefringence caused by the temperature change or the environmental vibration could be eliminated.
The reflected light is coupled back to the TE and TM polarizations in the IOCT. The light is directed to the phase analyzing section with the polarization controlling optical components. At the Y-branch, the light is divided into upper and lower paths. In the upper path, the phase modulator adjusts the initial phase difference between the light waves to operate the interferometer at the maximum sensitivity operating point, and the polarizer then filters out the TM component. In the lower path, the HWP is inserted for converting TE to TM, so that the polarizer installed in the lower path actually filters the TE component of the reflected light. Consequently, TE component passes through the upper path and TM component through the lower path, and they interfere at DC2. Depending on the amount of the nonreciprocal phase change resulting from the Faraday effect, the output signal will exhibit an amplitude modulation.
In the original reflection interferometry, for the compensation of the initial phase difference between the TE and TM polarizations, the reflected light should be propagating through the same path as the incident light [11]. In this case, however, because the light travels back through the exactly same path, it is difficult to apply the phase bias for adjusting the operating point of the sensor. Therefore, additional optical components such as a fiber-optic delay line and an electro-optic polarization-dependent phase modulator were mandatory [1]. On the contrary, in the proposed IOCT, the reflected light is directed into another path after DC1, and therefore, a low-speed thermo-optic phase modulator could be used to adjust the phase bias. Moreover, compared to the fiber-optic version of the PRRI, the path length of light in the waveguide device can be precisely controlled in the photomask layout. Hence, even when the reflected light is directed along another path, the path-length difference between forward and backward propagations could be minimized.
3. Fabrication and Characterization of the PRRI Chips
The functional devices integrated for PRRI chip are designed to be compatible with each other so that further modification of the fabrication process was not necessary. The polymer device has advantages on its simple fabrication process compared to other waveguide devices. The outlines of fabrication procedures are shown in Fig. 2 . On a silicon wafer, a fluorinated acrylate polymer ZPU available from ChemOptics, Co. is spin-coated and cured by UV light to form a cladding layer with a refractive index of 1.430. The polymer is further cross-linked through hard baking at 160°C for 1 h to increase long-term stability. The pattern of the waveguide core is formed by conventional photolithography and successive oxygen plasma etching with an etch rate of 1.0 μm/min. The core layer polymer with a refractive index of 1.440 is spin-coated over the waveguide pattern and cured to form an inverted rib structure. The inverted waveguide core has dimensions of 6.0 × 5.7 μm2 with an etch depth of 3.8 μm. Over the core layer, the first layer of the upper cladding is formed with a thickness of 3.1 μm.
Then, Au/Cr metal pattern with a length of 9 mm is formed as a polarizer for absorbing TM mode light through surface plasmon coupling. Another layer of the upper cladding is formed to cover the polarizer pattern. The total cladding thickness finally becomes 10 μm. On the surface of the cladding, a heating electrode of 40 μm wide and 3 mm long is formed using Cr-Au metal with a thickness of 10 nm and 100 nm, respectively, which results in a resistance of 48.6 Ω. On a 4-inch wafer, 30 devices are fabricated as shown in Fig. 3 .
Fig. 3 Photograph of the IOCT chip fabricated on a 4 inch silicon wafer and a single IOCT chip after dicing.
To evaluate the uniformity of devices fabricated on a wafer, the splitting ratios of a group of DCs were measured as shown in Fig. 4 . The DCs have coupling lengths from 900 μm to 1100 μm increasing in steps of 50 μm. It is shown that the experimental values lie on the line obtained by the simulation of beam propagation method. For two batches of samples, the results exhibited a slight deviation in the splitting ratios.
Fig. 4 Splitting ratios of 3-dB directional couplers obtained from two batches of the device fabrication; the experimental results well correspond to the BPM simulation results.
The waveguide polarizer with a thin metal layer absorbs TM mode light due to the surface plasmon mode coupling, while TE mode light is un-absorbed. The distance between the metal film and the waveguide determined as 1.8 μm for the strong absorption of TM mode light as well as a negligible loss of TE. Polarization extinction ratios of over 25 dB were typically observed in the fabricated device.
The HWP inserted in the lower path of the interferometer is made of a thin polymer film with the optic axis inclined at 45° to the substrate surface. Reactive mesogen was used to fabricate the birefringent polymer film. The detailed experimental results will be found elsewhere [12]. The HWP should be thin enough to reduce the diffraction loss when it is inserted in the middle of the waveguide. For inserting the film, a groove of 30 μm was formed by mechanical dicing. During the film insertion, the output polarization states were monitored on a Poincaré sphere as shown in Fig. 5 . Initially the output polarization was adjusted to the point of TM polarization by using a fiber-optic polarization controller. Then, by inserting the HWP, the point on the Poincaré sphere suddenly jumped to TE polarization point, indicating that polarization conversion had occurred. The plate was fixed by using a UV-curable epoxy. A slight deviation in the output polarization state occurred as one can observe from the Poincaré sphere. The deviation was within 3° in terms of the polarization angle.
Fig. 5 Polarization conversion shown on the Poincaré sphere during the fabrication of the TE/TM polarization converter.
For the assessment of the integrated device, where the individual devices could not be measured because the devices are connected to each other, we have developed a procedure of device characterization. Before constructing the OCT, the device performance was evaluated by operating the integrated phase modulator, and the interference signals for various input polarization states were observed. In this experiment, the input light was excited as the returned light from the fiber coil through the port supposed to be connected to output PM fiber. Then, an amplitude modulated signal depending on the phase modulation signal (Fig. 6(a) ) was observed at PD1 and PD2. Before inserting the HWP, for the TE polarized input light, the Mach-Zehnder interferometer exhibited an extinction ratio of 23.6 dB, as shown in Fig. 6(b). However, after the HWP is inserted, TE mode light cannot follow the lower path of the interferometer because of polarization conversion and subsequent polarization filtering. Consequently, interference could not occur, and the modulation signal disappeared as Fig. 6(c). Similarly, TM mode light could not follow the upper path, which results in a flat signal, as shown in Fig. 6(d). The slight modulation was a result of imperfect polarization conversion at the HWP or an unexpected polarization conversion at the phase modulator when a large heating power was applied. Similar to the operation of an actual OCT, an input light with a polarization angle of 45° was generated to excite both TE and TM modes. In this case, TE mode light follows the upper path, while TM mode light follows the lower path after the polarization conversion. The two field components interfered at DC2 and an amplitude modulated interference signal was observed as shown in Fig. 6(e). The extinction ratio of the signal was 17.6 dB, which was slightly less than the initial value before the insertion of the HWP. This value accumulates the effects of all of the components integrated on the IOCT chip. The phase retardation and the optic axis angle of the inserted HWP exhibited most sensitive effect on the final extinction ratio.
Fig. 6 Optical interference signals depending on the input polarization states produced by modulating the PM with (a) a 10-Hz triangular source signal. Before the HWP insertion, for TE polarization, the interference signal was observed as shown in (b). After the HWP insertion, no modulation was observed for (c) TE polarization, or (d) TM polarization, while a strong interference signal was observed for 45° polarized light (e).
4. Current Sensing Experiment
The IOCT chip was incorporated in the PRRI current sensor along with the SLD light source and the fiber sensing coil. The SLD used in this study had a 3-dB bandwidth of 60 nm, which corresponds to the coherence length of 40 μm. The fiber sensing module included a fiber-optic QWP, an annealed fiber coil, and a fiber-optic mirror. The QWP was fabricated by using a piece of PM fiber with a specific length for introducing a phase shift of λ/4 between the fast and slow axis components. Then, the PM fiber piece was spliced on the long PM fiber with an angle of 45°.
The fiber coil was prepared by winding a bare fiber 10 times around a ceramic frame, and annealing the frame at 850°C for 24 h. The annealing was carried out to eliminate the bending-induced birefringence. The fiber-optic mirror had a Cr-Au metal coating on one side, and it was attached to the annealed fiber coil at the other side. Detailed fabrication procedures of the fiber-optic components are described in a previous publication of the authors [9].
The reflected signal from the sensor coil was detected by using two photodetectors connected to the PRRI chip. The phase difference between the fast and slow axis components was converted into an optical-amplitude change after the interference at DC2. Then the two output signals of DC2 were compared to find the difference and normalized by the total power to reduce the source power fluctuation noise. By applying a bias voltage to the phase modulator of IOCT chip, the operating point of the interferometer was maintained to have a phase difference of π/2, at which the maximum signal amplitude was produced. The optical signal probed at the photodetector is shown in Fig. 7 . To demonstrate the large bandwidth of the OCT, a rectangular input signal of 2.5 kHz was used as a current source. The input current waveform was measured by using a high frequency current sensing probe (TCP 404XL, Tektronix). As one can see, the output optical signal duplicated the high-frequency component of the rectangular signal with no considerable frequency filtering. The maximum input current was limited to 400 A by the source capacity. Appropriate electrical signal processing was used to calculate the amplitude of the optical interference signal as final sensor output. With respect to the input current intensity, the optical sensor output was obtained as shown in Fig. 8 . For the input current of 400 A, the modulation depth of the signal corresponded to the phase change of 1.44° resulting from the Faraday effect. The output signal exhibited excellent linearity with a measurement accuracy of 0.2%, which is suitable for the power metering application in a smart-grid power network.
Fig. 7 Optical output signal of the OCT responding to an input current of 2.5-kHz rectangular waveform.
Fig. 8 The final output of the OCT after the signal processing, where the deviation of the output signal from the linear characteristics was within 0.2%.
5. Conclusion
An integrated-optic version of the optical current transducer was constructed by integrating various polymeric waveguide devices, which include polarization maintaining 3-dB couplers, TE-pass polarizers, TE/TM polarization converters, and thermo-optic phase modulators. The unique method of fabricating a polymer waveguide device facilitated the integration of the various functional devices on a single substrate without additional modification of the fabrication process. Compared to the fiber-optic current sensors, the integrated optic devices exhibited the merits of precise control of optical path length and polarization stability across the components. In the current sensor experiment, the output signal exhibited excellent linearity with a measurement accuracy of 0.2%, which satisfies the requirement of current sensors for power metering. Optical temperature sensor has slight temperature dependence due to the temperature dependence of the Faraday effect, and it could be compensated by incorporating a phase retarder with an initial off-set.
Acknowledgement
This research was supported by the Korea Science and Engineering Foundation (KOSEF) grant (2009-0079553) and the World Class University Program through the National Research Foundation of Korea (R31-2008-000-20004-0), Ministry of Education, Science and Technology, Korea.
References and links
1. K. Bohnert, P. Gabus, J. Nehring, and H. Brandle, “Temperature and vibration insensitive fiber-optic current sensor,” J. Lightwave Technol. 20(2), 267–276 (2002). [CrossRef]
2. F. Rahmatian, and J. N. Blake, “High-voltage fiber optic current sensors,” in Proc. IEEE PES General Meeting, Montreal, Quebec, (2006), pp. 1129.
3. M. Hino, S. Hase, K. Ajiki, and M. Akagi, “Optical fiber current transformer applications on railway electric power supply systems,” Quarterly Rept. Railw. Tech. Res. Inst. 45, 59–63 (2004).
4. J. D. P. Hrabliuk, “Optical current sensors eliminate CT saturation,” in Proc. 2002 IEEE PES Winter Meeting, 2, pp. 1478–1481.
5. K. Bohnert, P. Gabus, J. Nehring, H. Brandle, and M. Brunzel, “Fiber-optic current sensor for electrowinning of metals,” J. Lightwave Technol. 25(11), 3602–3609 (2007). [CrossRef]
6. M.-C. Oh, M.-H. Lee, and H.-J. Lee, “TE-pass and TM-pass waveguide polarizers with buried birefringent polymer,” Electron. Lett. 35(6), 471–472 (1999). [CrossRef]
7. M.-C. Oh, M.-H. Lee, and H.-J. Lee, “Polymeric waveguide polarization splitter with a buried birefringent polymer,” IEEE Photon. Technol. Lett. 11(9), 1144–1146 (1999). [CrossRef]
8. M.-C. Oh, S.-Y. Shin, W.-Y. Hwang, and J.-J. Kim, “Wavelength insensitive passive polarization converter fabricated by poled polymer waveguides,” Appl. Phys. Lett. 67(13), 1821–1823 (1995). [CrossRef]
9. M.-C. Oh, J.-K. Seo, K.-J. Kim, H. Kim, J.-W. Kim, and W.-S. Chu, “Optical current sensors consisting of polymeric waveguide components,” J. Lightwave Technol. 28(12), 1851–1857 (2010). [CrossRef]
10. V. Minier, D. Persegol, J. L. Lovato, and A. Kevorkian, “Integrated Optical Current Sensor with low-birefringence optical waveguides”, 12th Int. Conf. on Optical Fiber Sensors, 1997 OSA Technical Digest Series, 16, 104 – 107, (1997).
11. A. Enokihara, M. Izutsu, and T. Sueta, “Optical fiber sensors using the method of polarization-rotated reflection,” J. Lightwave Technol. 5(11), 1584–1590 (1987). [CrossRef]
12. W.-S. Chu, S.-M. Kim, J.-W. Kim, K.-J. Kim, and M.-C. Oh, “Waveguide polarization converters incorporating UV-curable reactive mesogen,” IEEE Photon. Technol. Lett. submitted.
