August 1, 2009 / Vol. 34, No. 15 / OPTICS LETTERS
2369
Adiabatically tapered splice for selective excitation of the fundamental mode in a multimode fiber Yongmin Jung,1,* Yoonchan Jeong,1,2 Gilberto Brambilla,1 and David J. Richardson1 1
Optoelectronics Research Centre, University of Southampton, Southampton, SO17 1BJ, UK 2
[email protected] *Corresponding author:
[email protected]
Received April 28, 2009; revised June 25, 2009; accepted July 4, 2009; posted July 9, 2009 (Doc. ID 110701); published July 31, 2009 We propose a simple and effective method to selectively excite the fundamental mode of a multimode fiber by adiabatically tapering a fusion splice to a single-mode fiber. We experimentally demonstrate the method by adiabatically tapering splice (taper waist= 15 m, uniform length= 40 mm) between single-mode and multimode fiber and show that it provides a successful mode conversion/connection and allows for almost perfect fundamental mode excitation in the multimode fiber. Excellent beam quality 共M2 ⬃ 1.08兲 was achieved with low loss and high environmental stability. © 2009 Optical Society of America OCIS codes: 230.1150, 060.2340, 060.3510.
The robust excitation of just the fundamental mode in a multimode fiber is a significant issue in the design and implementation of many high-performance fiber lasers, optical sensors, and other photonic devices. For example, optimization of the seed-launch conditions in high-power fiber lasers that are normally based on large-core multimode fibers is extremely important to obtain high beam quality [1]. In a multimode fiber Bragg grating, the restricted mode excitation can be used to generate a single rather than an otherwise multipeaked reflectivity spectrum [2]. To achieve near-single-mode performance in multimode fibers, a mode-selective filtering technique (e.g., coiling the fiber and exploiting the greater bend loss sensitivity of higher order modes) is generally required in conjunction with careful adjustment of the launch conditions in order to suppress higher-order modes [3–5]. However, this often involves the use of bulk optic components and/or tedious, timeconsuming alignment processes. In this Letter we propose a scheme for fundamental mode excitation and solve the delicate mode launching problem by using an adiabatically tapered splice between two dissimilar fibers. This tapered splice technique has previously been used to reduce the splice loss between two dissimilar single-mode fibers [6], to couple optical power from a single-core fiber to a multicore fiber [7,8], and to realize low-loss 1 ⫻ 2 Y-junctions [9,10]. However, to our knowledge, to date there has been no report on the application of single-mode excitation in multimode fibers utilizing this approach. Here, we experimentally demonstrate that successive core-cladding mode conversion in the tapered splice can lead to almost perfect fundamental mode excitation in a multimode fiber. Figure 1 represents an idealized tapered splice configuration for exciting the fundamental mode in a multimode fiber, which can be fabricated by fusion splicing a single-mode fiber (SMF) to a multimode fiber (MMF) and adiabatically tapering the splice point. As depicted in the evolution of the guided mode (Fig. 1, top), the fundamental core mode of the SMF core共SMF兲 兲 is continuously mode coninput fiber 共LP01 0146-9592/09/152369-3/$15.00
clad共SMF兲 verted into a guided cladding mode 共LP01 兲 in clad共SMF兲 the waist region by the down-taper. The LP01 cladding mode further propagates along the splice clad共SMF兲 joint with negligible optical loss 共LP01 clad共MMF兲 → LP01 兲 and is then coupled back into a guided core共MMF兲 兲 by core mode in the MMF output fiber 共LP01 the up-taper [11,12]. Note that the cladding modes are defined by the cladding/air interface and that the original core has a negligible effect on their guidance. Therefore, pure fundamental excitation in a multimode fiber with minimal optical losses can be core共SMF兲 achieved by successive mode evolution 共LP01 clad共SMF兲 clad共MMF兲 core共MMF兲 → LP01 → LP01 → LP01 兲 along the adiabatically tapered splice. As a simple example, a standard telecom fiber (Corning SMF-28) was spliced to a multimode fiber (core diameter= 26 m, NA= 0.1) that supports up to four modes at a wavelength = 1.55 m. After fusion splicing the two fibers, the splice point was positioned at the center of a taper rig and adiabatically tapered using the modified “flame-brushing” technique [13]. The resulting tapered splice has a uniform waist diameter of 15 m and a length of 40 mm. In adiabatic tapers the changes of the taper angle are so slow that little if any mode conversion takes place. To investigate the modal guidance, the transmission spectra of the tapered splice were recorded during fabrication using an incoherent white-light source and an optical spectrum analyzer (see Fig. 2).
Fig. 1. (Color online) Schematic of an adiabatically tapered splice for exciting the fundamental mode in an MMF. © 2009 Optical Society of America
2370
OPTICS LETTERS / Vol. 34, No. 15 / August 1, 2009
Fig. 2. (Color online) Transmission spectra of the MMF and the proposed adiabatic tapered splice both before/after tapering.
In the case of an MMF (dotted curve in Fig. 2) with a large number of core modes, the transmission spectrum shows a much higher power level than an SMF, and the wavelength dependence is not resolved. However, when the SMF was simply spliced to the input end of the MMF (dashed curve in Fig. 2), the transmission spectrum of the composite structure became similar to that of the conventional SMF where higher-order mode cutoffs can be observed due to the guidance properties of the input SMF. After adiabatically tapering the splice point (solid curve in Fig. 2), intermodal interference appears in the 400–1250 nm spectral range, while no perturbation occurs above 1250 nm. Note that for the tapered splice there is no significant spectral change in the single-mode operation range of the input fiber, and the optical loss is negligible (⬍0.1 dB at = 1.55 m). More detailed modal guidance properties and spectral tendencies are described in [11]. To confirm stable single-mode guidance of the SMF–MMF tapered splice, the far-field pattern was imaged at 1550 nm using a 50⫻ microscope lens and a CCD camera. First of all, we launched the coherent laser light directly into an MMF and imaged the output field. As shown in Fig. 3(a), a severe speckle pattern was observed, resulting from the superposition of various fiber modes arriving at the output fiber end face with different phases. The resulting interference pattern was highly sensitive to external perturbations such as bending. However, by splicing an SMF to the input end of the MMF (SMF–MMF splice), the measured speckle noise [Fig. 3(b)] was substantially reduced owing to the restricted mode excitation from the on-axis underfilled launch into the MMF. Here, coherent laser light 共 = 1.55 m兲 was injected into the SMF input fiber, and the farfield patterns were observed at the output end of the MMF. However, the light propagating from an SMF into an MMF still excites several higher-order modes in the MMF and generates nonnegligible intermodal interference in the far-field patterns. Finally, using the adiabatic tapered splice, as shown in Fig. 3(c), clean excitation of the fundamental mode was suc-
Fig. 3. (Color online) Measured far-field intensity profile: (a) launching light into the MMF (b) without and (c) with an adiabatic tapered splice.
cessfully obtained within the MMFs as evidenced by the high-quality single-mode output beam. To further investigate the stability of the single-mode operation external bends (1 or 2 turns with 30 mm bend diameter) were applied to the output MMF, but no degradation was observed in either the mode profile or the transmission spectrum. Since no other modes are launched into the MMF except the fundamental modes, the bending or twisting of the MMF does not affect the fundamental mode pattern along the fiber. To quantitatively analyze the spatial beam quality, the beam quality factor 共M2兲 was measured using a CCD camera-based method [14,15], and the result is shown in Fig. 4. Generally, the propagation of a Gaussian beam can be fully specified by either its beam waist diameter or its far-field divergence. In our experiment, the output beam from the end of the fiber was collimated using a 20⫻ objective lens and
Fig. 4. (Color online) Measured beam quality factor 共M2兲 at the output end of the MMF.
August 1, 2009 / Vol. 34, No. 15 / OPTICS LETTERS
then focused using an antireflection-coated planoconvex lens (focal length= 200 mm). A CCD camera mounted on a translation stage and beam profiling software (Spiricon LBA-710PC) was used to capture and analyze the beam profiles. In Fig. 4, it can be noticed that the beam waist and far-field divergence obtained from the MMF itself are several times larger than those of the SMF–MMF tapered splice. According to the hyperbolic fit to the measured beam width, the optical beam quality factors for the MMF, SMF– MMF splice, and SMF–MMF tapered splice were estimated to be 3.03, 2.11, and 1.08, respectively. As predicted from the far-field patterns in Fig. 3, the SMF–MMF splice provides improved beam quality owing to the on-axis, underfilled launch condition, and the SMF–MMF tapered splice performs the function of launching the light from the source exclusively into the fundamental mode in an MMF, confirming that close to diffraction-limited performance 共M2 ⬃ 1.08兲 was obtained. In conclusion, efficient and robust fundamentalmode excitation in a multimode fiber was achieved by using an adiabatically tapered splice. Through the core共SMF兲 successive mode transformation 共LP01 clad共SMF兲 clad共MMF兲 core共MMF兲 → LP01 → LP01 → LP01 兲, a high2 quality beam 共M ⬃ 1.08兲 was achieved with low optical loss at a wavelength of 1.55 m. We expect that the proposed scheme can be directly applicable to other operational wavelength ranges and will provide a new degree of freedom in the design of highperformance fiber lasers, optical fiber sensor systems, and other photonic devices that are based on largecore, multimode fibers.
2371
The authors thank the Engineering and Physical Sciences Research Council UK (EPSRC) for financial support; G. B. gratefully acknowledges the Royal Society (London, UK) for his Research Fellowship. References 1. Y. Jeong, J. K. Sahu, D. B. S. Soh, C. A. Codemard, and J. Nilsson, Opt. Lett. 30, 2997 (2005). 2. T. Mizunami, T. V. Djambova, T. Niiho, and S. Gupta, J. Lightwave Technol. 18, 230 (2000). 3. M. E. Fermann, Opt. Lett. 23, 52 (1998). 4. Y. Jeong, J. K. Sahu, D. N. Payne, and J. Nilsson, Opt. Express 12, 6088 (2004). 5. S. G. Leon-Saval, T. A. Birks, J. Band-Hawthorn, and M. Englund, Opt. Lett. 30, 2545 (2005). 6. D. B. Mortimore and J. V. Wright, Electron. Lett. 22, 318 (1986). 7. L. Yuan, Z. Liu, J. Yang, and C. Guan, Appl. Opt. 47, 3307 (2008). 8. L. Yuan, Z. Liu, and J. Yang, Opt. Lett. 31, 3237 (2006). 9. N. Healy, E. McDaid, D. G. Murphy, C. D. Hussey, and T. A. Birks, Electron. Lett. 42, 740 (2006). 10. J. D. Minelly and C. D. Hussey, Electron. Lett. 23, 1087 (1987). 11. Y. Jung, G. Brambilla, and D. J. Richardson, Opt. Express 16, 14661 (2008). 12. Y. Jung, G. Brambilla, and D. J. Richardson, Opt. Express 17, 5273 (2009) 13. G. Brambilla, F. Koizumi, X. Feng, and D. J. Richardson, Electron. Lett. 41, 400 (2005). 14. International Organization for Standardization (ISO) Standard EN ISO 11146: 2000, (ISO, Geneva, Switzerland, 2000). 15. J. D. Shephard, P. J. Roberts, J. D. C. Jones, J. C. Knight, and D. P. Hand, J. Lightwave Technol. 24, 3761 (2006).
