Tunability of optical components has been one of the central issues in optical communications and sensing, and band rejection filters (BRFs) would most benefit from the wide-tuning capability. Recent development of low-water-peak, single-mode fibers (SMFs) [1] has expanded the optical communication windows into O ($1260–1360\mathrm{nm}$), E ($1360–1460\mathrm{nm}$), S ($1460–1530\mathrm{nm}$), C ($1530–1560\mathrm{nm}$), L ($1560–1625\mathrm{nm}$), and U ($1625–1675\mathrm{nm}$) bands [2]. A single BRF that can cover the entire communication bands would be of high interest for seamless interband communications and signal processing. A long-period grating (LPG) that modulates the effective index of the fundamental core mode with a phase-matching period of $\sim 500\mathrm{\mu m}$, has been widely used as a compact fiber BRF, but its tuning ranges have been limited to within a single communication band [3]. Acousto-optic tunable filters (AOTFs) have shown wider tuning ranges, and a recent photonic crystal fiber (PCF) AOTF showed a tuning over $1000\mathrm{nm}$ [4]. Despite the wide tuning range, the PCF AOTF required a special dual-mode PCF and inherently suffered from a high insertion loss due to the mode mismatch between conventional SMFs and PCF. Therefore, all-fiber devices with a wide tuning range, low insertion loss, and high compatibility with conventional SMF would be of great interest.

Mechanically tunable coupling characteristics in fused taper fiber couplers (FTFCs) have been reported by Birks [5], and recently, the principle has been combined into a micro-optical waveguide (MOW) on a micro actuating platform (MAP) [6] by the authors for appli cation in various compact tunable fiber devices such as an interband router [7, 8], and a variable optical attenuator [9, 10].

In this Letter, we further extended a MOW on a MAP structure made of conventional SMF FTFCs to experimentally demonstrate a BRF whose tuning range cover the entire O, E, S, C, L, and U bands with a very low loss and high rejection efficiency, for the first time to the best knowledge of the authors.

As the fundamental ${\mathrm{LP}}_{01}$ mode in one of the input fibers of a $2\times 2$ FTFC enters the fused taper region, it excites both symmetric and antisymmetric modes in the taper to result in a mode beating. If the taper length is integer multiple times one-half of the beat length, a perfect power transfer occurs between two fiber arms [11]. When a torsional stress is locally applied over the fused region, the effective refractive index, $n_{\text{eff}}$, changes by the stress-optic effect as below [5]:

In this study, we proposed a compact FTFC-based MOW on MAP, and its schematic diagram is shown in Fig. 1. Here, the waist of an optimized $2\times 2$ FTFC, whose width is between 20 and $70\mathrm{\mu m}$, serves as the MOW and is mounted on a MAP that is an electrically controlled, high-precision rotational stage. When the light is coupled from the arm-2 to arm-1 in a certain spectral range, the transmission through arm-2 results in typical characteristics of a BRF, as schematically shown in Fig. 1. When a highly localized torsional stress with a certain pitch, b, is applied along the waist, the $n_{\text{eff}}$ changes, and subsequently, the peak wavelength of the coupling shifts to provide a spectral tuning capability.

In this study, we used an SMF, Corning HI1060, whose cutoff was near $920\mathrm{nm}$ so the proposed device could provide a significantly broader tuning wavelength range than prior reports based on conventional SMF with a longer cutoff wavelength. In prior flame brushing tech niques for coupler fabrication using hydrogen or propane fuel, OH-induced excess loss near $1380\mathrm{nm}$ has been the major source of excess loss [12], which has significantly hindered the E-band ($1360–1460\mathrm{nm}$) applications. In this study, fusion and tapering of optical fibers were all processed using a compact, electrical ceramic heater to fundamentally remove the OH accumulation during fabrication processes. The electrical heater provided a Gaussian heat zone of a few millimeters length with the maximum temperature of $1500°\mathrm{C}$. Evolution of the transmission spectra in one of the fiber arms of an FTFC were monitored in situ using a white light source and an optical spectrum analyzer.

Before we mounted the optimal FTFC MOW on the MAP, its pristine transmission spectrum showed a minimum at $1350\mathrm{nm}$ with more than $30\mathrm{dB}$ rejection efficiency, $3\mathrm{dB}$ bandwidth of $50\mathrm{nm}$, and an insertion loss less than $0.2\mathrm{dB}$. The spectrum is shown in plot 1 of Fig. 2.

The FTFC MOW with a waist diameter of $70\mathrm{\mu m}$ was carefully mounted on the MAP with a slight axial tension to prevent sagging. By precisely rotating the MAP, the torsional stress was locally focused over the fused waist. The rotation angle was varied by the increment of $60°$ with the precision of $\pm 0.1°$. At each rotation angle, transmission spectrum was measured, and results are summarized in Fig. 2. Note that we could not observe the OH-induced loss near $1380\mathrm{nm}$ in contrast to significant OH loss induced in oxyhydrogen flame brushing method 12, which enabled seamless tuning over the wide spectral range. As we increase the rotation angle, the band rejection peak shifted toward longer wavelengths, covering the O, E, S, C, L, and U bands sequentially. When the rotation angle exceeds $240°$, as in spectra 5 and 6 in Fig. 2, the transmission spectrum showed another coupling in the shorter wavelengths near 950 and $1060\mathrm{nm}$, respectively. This is the typical nature of FTFCs to show a cyclic coupling in the spectral domain. It is also noted that the peak of the rejection band monotonically shifts to a longer wavelength regardless of the rotation direction clockwise or counterclockwise. It was also observed that the FWHM and the peak rejection efficiency changed in a reproducible manner.

To further investigate and analyze the tuning mechanism of the proposed device, two types of FTFC MOWs were fabricated with a waist radius of $20\mathrm{\mu m}$ in addition to a $70\mathrm{\mu m}$ sample by adjusting the tapering conditions. The FTFC with the waist radius of $70\mathrm{\mu m}$ showed a range of spectral shift, $\sim 350\mathrm{nm}$ per one turn ($=360°$) with a parabolic response to the rotation angle, as shown in Fig. 3a. In contrast, the device with a FTFC waist radius of $20\mathrm{\mu m}$ showed a slightly wider spectral shift range, $\sim 370\mathrm{nm}$ per one turn, with an almost linear response whose slope was $\sim 1.3\mathrm{nm}/°$, as shown in Fig. 3b.

The differences in the spectral response of the device are attributed to the mechanical behavior of microsilica wire at different radii [13]. In general, a thinner silica wire can sustain a higher torsional stress per unit length to allow smaller pitch b as in Fig. 1. Experimentally, we could observe the helical pitch of several tens of micrometers on the FTFC MOW with a $20\mathrm{\mu m}$ waist. In contrast, the pitch in the TFTC MOW with a $70\mathrm{\mu m}$ waist was 1 order of magnitude longer owing to its higher stiffness. These helical pitch differences under the torsional stress are mainly attributed to the different behav ior of the devices, as shown in Fig. 3. The spectral shift in the proposed device showed a strong dependence on the FTFC MOW waist, which could provide a new degree of freedom to design both the operating spectral range and tuning response of the device.

The variations of the FWHM, equivalently $3\mathrm{dB}$ bandwidth, of the band rejection efficiency were also measured. The experimental results are summarized in Fig. 3b. Here we measured the spectral width in nano meters, where the band rejection efficiency takes a value lower than the peak by $3\mathrm{dB}$. The FWHM was plotted as a function of rotation angle in Fig. 3. The FTFC MOW of a $70\mathrm{\mu m}$ waist showed FWHM in the range from 22 to $40\mathrm{nm}$ with its minimum at the initial position near $1357\mathrm{nm}$, as shown in Fig. 3b. The cyclic peaks appeared in the shorter wavelengths beyond the rotation angle of $240°$ and showed a FWHM of $30–35\mathrm{nm}$. In the case of a $20\mathrm{\mu m}$ waist, the FWHM varied in the range from 22 to $71.5\mathrm{nm}$, as shown in Fig. 3b. The FWHM generally increased with the rotation angle, which is also attributed to the changes in the coupling constants by photo-elastic effects in the microsilica wire [14].

Detailed variations in the maximum band rejection efficiency are summarized in Fig. 3b. The maximum band rejection efficiency changed from 32.5 to $11\mathrm{dB}$ as the rotation angle increased from 0 to $300°$ for the proposed device with a FTFC MOW waist diameter of $70\mathrm{\mu m}$. The shorter wavelength peaks beyond the rotation angle of $240°$ showed a band rejection efficiency of $\sim 22\mathrm{dB}$. The range of band rejection efficiency was $17–25\mathrm{dB}$ for the case of $20\mathrm{\mu m}$ waist.

The tuning process in the proposed device was found to be reversible and reproducible such that, if the torsion is released, the FTFC MOW returns to its original characteristics. The device with an FTFC MOW of a $20\mathrm{\mu m}$ waist radius provided more than six turns and three turns in the case of a $70\mathrm{\mu m}$ waist radius, within which the proposed device did not show any sign of hysteresis in the transmission spectra. More detailed investigation on the device’s mechanical reliability is being pursued by the authors.

In conclusion, we successfully demonstrated a tunable BRF made of FTFC MOW mounted on a MAP by applying torsional stress over the waist of an FTFC MOW. Torsional stress induced a systematic change in the effective index in the waist to provide effective tuning in a wide spectral range. The peak band rejection efficiency was more than $25\mathrm{dB}$ and tunable from a 1060 to $1670\mathrm{nm}$ covering. The insertion loss was kept below $0.2\mathrm{dB}$, and polarization-dependent loss was below $0.1\mathrm{dB}$. Two types of FTFCs were fabricated with the waist radius of 20 and $70\mathrm{\mu m}$. Depending on the waist diameter, the device showed different tuning characteristics in terms of the spectral shift, FWHM, and the peak band rejection efficiency. In a laboratory environment, the tuning process was reproducible without any hysteresis. The proposed device can find ample potential in tunable BRF for ultrawide, band optical communications and sensing applications.

This work was supported in part by the Brain Korea 21 Project, in part by the National Research Foundation grant funded by the Ministry of Education, Science, and Technology (MEST) (grants 2010-0018442, 2009-00479 EC-FP7/2007-2013 219299 GOSPEL, R15-2004-024-00000-0, F01-2009-000-10200-0, and 2009-00541), in part by the Korea Institute of Industrial Technology Evaluation and Planning (grants 2009-8-0809 and 2010-8-1415), and in part by the Agency for Defense Development (grant 2010-8-1806).

 

Fig. 1 Schematic diagram of the proposed device. The waist of an optimized $2\times 2$ fused taper fiber coupler is mounted on a rotational MAP, where electrically controlled torsional stress is applied.

Download Full Size | PDF

 

Fig. 2 Changes in the transmission spectra of the proposed device under torsional stress. The integer, n, assigned under each spectrum, indicates the rotation angle of $(n-1)\times 60°$. The spectral ranges for O, E, S, C, L, and U bands are overlaid. Here, the waist of the FTFC is $70\mathrm{\mu m}$.

Download Full Size | PDF

 

Fig. 3 (a) Experimental measurements for the spectral shifts for FTFC of waist widths 70 and $20\mathrm{\mu m}$. The reference peak wavelength is $1350\mathrm{nm}$ as in plot 1 of Fig. 2. The cyclic coupling in the shorter wavelength is illustrated as the negative shift. (b) Changes in the FWHM and bend rejection efficiency under torsional stress that were applied by rotating the MAP by increments of $60°$. FWHM and rejection efficiency of the band rejection peaks within the tuning range for the FTTC MOW waist diameters of 70 and $20\mathrm{\mu m}$.

Download Full Size | PDF

1. G. A. Thomas, B. I. Shraiman, P. F. Glodis, and M. J. Stephen, Nature 404, 262 (2000). [CrossRef]   [PubMed]  

2. International Telecomunication Union (ITU), “$100\mathrm{GHz}$, $50\mathrm{GHz}$, $25\mathrm{GHz}$ channel spacing (G.694.1),” http://www.itu.int/.

3. A. M. Vengsarkar, P. J. Lemaire, J. B. Judkins, V. Bhatia, T. Erdogan, and J. E. Sip, J. Lightwave Technol. 14, 58 (1996). [CrossRef]  

4. K. S. Hong, H. C. Park, B. Y. Kim, I. K. Hwang, W. Jin, J. Ju, and D. I. Yeom, Appl. Phys. Lett. 92, 031110 (2008). [CrossRef]  

5. T. A. Birks, Appl. Opt. 28, 4226 (1989). [CrossRef]   [PubMed]  

6. W. Shin and K. Oh, Opt. Express 12, 4378 (2004). [CrossRef]   [PubMed]  

7. J. B. Eom, H.-R. Lim, K. S. Park, and B. Lee, Opt. Lett. 35, 2726 (2010). [CrossRef]   [PubMed]  

8. W. Shin, Y. Jeong, J. Lee, and K. Oh, J. Lightwave Technol., 23, 533 (2005). [CrossRef]  

9. M. B. J. Diemeer, W. J. De Vries, and K. W. Benoist, Electron. Lett. 24, 457 (1988). [CrossRef]  

10. T. A. Birks, Electron. Lett. 27, 1087 (1991). [CrossRef]  

11. N.-K. Chen, C.-L. Lee, and S. Chi, Opt. Express 15, 17747 (2007). [CrossRef]   [PubMed]  

12. W. Shin, U. C. Ryu, and K. Oh, Electron. Lett. 38, 214 (2002). [CrossRef]  

13. J. F. Doyle and J. W. Phillips, Manual on Experimental Stress Analysis (Society for Experimental Mechanics, 1989).

14. C. N. Alexeyev, A. V. Volyar, and M. A. Yavorsky, J. Opt. A 10, 095007 (2008). [CrossRef]