We demonstrate a very simple all-optical switch (AOS) based on the bend effect of a nm fiber taper driven by outgoing light. The AOS operates at relative low control power compared with those of using nonlinear effect. In our configuration, we find that the turnoff and turnon time of the AOS are about 500ms and 760ms, respectively. The optimized extinction ratio of about 15db under 32mW control light at 980nm is achieved. This AOS effect will be very useful for creation of different interesting functional devices.
©2009 Optical Society of America
Recently, Tong and co-workers showed a two-step method to draw a commercial standard single-mode fiber down to 50nm diameter by alcohol burner [1, 2]. B.J. Li and co-workers fabricated nanofiber with a diameter down to 60nm and a length up to 500nm by drawing from molten poly(trimethylene terephthalate) [3, 4]. Due to their smooth surface, low loss, strong light confinement and evanescent wave , such nanofibers are potential for a variety of compact optical devices. Various kind of devices based on the micro/nano fiber were demonstrated, which included resonators [6–9], micro-lasers [10, 11], add-drop filters , Mach-Zehnder interferometer , nanofiber sensors [14, 15], compact couplers . To the best of our knowledge, AOS based on the nm fiber taper has never been reported.
As is well known, conventional all-optical switch is based on third order optical nonlinear effect . But unfortunately the third order nonlinear coefficients are very small  and therefore very high power of control light or large size of the devices are required, which limit the implement of small-size all-optical switch (AOS) operated at low power, although enhancement of intensity of light can be achieved in nanofibers. In order to solve the problem of high operation power, several novel mechanisms have been proposed and demonstrated, which include relative index change of photonic crystal by light-induced capillary condensation , disorder-to-order phase transition of polymer spheres , and quantum-interference [20, 21]. Very recently, an interesting effect, that a nm silica filament in a sealed glass chamber is driven by optical force related to Abraham momentum when light emerges from the filament end, is reported . In this paper, we show experimentally that this effect can be used to implement a very simple low-operated-power AOS.
2. Structure of AOS and experimental setup
The structure and the experiment setup of the novel type of AOS are showed in Fig. 1. The AOS consists of two fiber tapers mounted in an erect sealed glass chamber [Fig. 1(c)]. The distance between two ends of tapers can be adjusted by a three dimensional (3D) stage. The taper above is smaller than that below. The image shown in Fig. 1(c) is taken by a digital camera (Canon G8) at maximum zoom-in mode, after a quartz plane-convex lens with 3.8mm focus length. In experiment, a signal light at 650nm from a semiconductor laser with 2Hz repetition and a control light at 980nm from a power-tunable (0-120mW) semiconductor laser are combined by a wavelength-division-multiplexer (WDM1) and then sent into the smaller taper through the single mode fiber. Before turning on the control light, the larger taper is well collimated by 3D stage so that it is coaxial with the smaller one. In this case, the 650nm light emerging from the end-face of the smaller taper is mostly coupled into the larger one. The coupled-in signal light passes through the output fiber and then goes to the second WDM (WDM2). WDM2 is used to filter out the control light and let the signal light travel to a photomultiplier when the control light is turned on. The photomultiplier is connected with an oscilloscope (Tektronix TDS3032) for observing the influence of the control light on the signal. It should be noticed that, to avoid the influence of the flowing air on the smaller taper, in our exploring experiment, two tapers are sealed in a glass chamber with a diameter of 10cm. In fact, the chamber can be reduced to a very small size and the size of AOS is principally dependent on both the length and maximum displacement of the smaller taper.
The two tapers used are fabricated by using flame-heated taper-drawing technique, which is a little different from that reported in reference . We use two 1D stages, instead of hands, for drawing a standard single mode fiber (Corning SM28) softened by a homemade alcohol torch with a nozzle of about 5mm diameter. The drawing speed, which is an important parameter in determining both size and length of fiber tapers, can be easily controlled by the stages. At the beginning, the drawing speed is about 10μm/s, and is increased gradually to about 100μm/s when the taper becomes small enough (about 5μm diameter). Thus the tapers can satisfy the adiabatic condition and has low loss. The loss of the tapers measured at 650nm is lower than 0.024db/mm. The tips of the smaller and the larger taper are respectively shown in Figs. 1(a) and 1(b), which are taken by a microscope (LEICA DM2500P) with 50× objective lens and high performance CCD. Figure 1(a) shows the uniform diameter of the smaller taper near its end, about 500nm. In contrast, the diameter of the larger taper showed in Fig. 1(b) decreases from 5.0μm to 2.7μm/s along 34μm length. It should be noticed that to make the AOS work, the space between two tips should not be too large, which is less than 500nm in our case, because it is found that, for a space larger than 500nm the light emerging from the smaller taper spreads out very quickly due to the diffraction and very little light passes through the AOS.
3. Experiment of AOS
Firstly, we study the extinction ratio of such type of AOS with different power of control light. Figure 2 shows different wave shapes of the signal captured by oscilloscope for different control powers. When without control light, the two tapers are coaxial as showed in Fig. 1(a) and the 650nm light mostly passes through the AOS. The signal reaches its maximum as shown in Fig. 2(a). When the control light is turned on, the smaller taper is bent by the optical force. It is observed in experiment that the higher the power of control light is, the more bent the smaller taper becomes. The power (about 0.5mW) of 650nm light used in the experiment is too weak to drive the taper. So its effect can be ignored. We find that, at the beginning, the signal does decrease monotonically with the increase of the bending of the smaller taper [see Figs. 2(b) and 2(c)], but the situation is inversed when the degree of bend exceeds a certain value [see Figs. 2(d) 2(e) and 2(f)]. This is because the more bent the smaller taper becomes, the more the leakage of 650nm light from the flank of the smaller taper appears, which partly enters the larger taper. Hence, when the control power exceeds some value, the extinction ratio of AOS will not be directly proportional to the power of control light and there exists an optimized value. From Figs. 2(b)–2(f), we can calculate the extinction ratios of the AOS, which are about 4.5db, 8.9db, 15db, 12db, and 10db, corresponding to the control light of 21.9mW, 29.9mW, 32mW, 34mW, and 42.3mW, respectively. One see that the optimized extinction for present AOS is 15db, corresponding to a control light of 32mW.
To observe the detail of turn -on and -off of AOS clearly, for the experiment shown in Fig. 3 (the same Fig. 5 below), we used a voltage applied on the photomultiplier as four times as that in Fig. 2. Figure 3 shows the wave shape of signal captured by oscilloscope when control light is turned on suddenly and the AOS is becoming “off”. When the control light is turned on, the signal decays to 5% in a time of about 500ms as indicated in Fig. 3. To show how the AOS works, here we display video frames [Figs. 4(a)–4(h)] of the moving smaller taper as the result of the optical force taken by a digital camera (Canon G8) at movies mode with a rate of 15 frames/s. Figure 4(a) shows that, the smaller taper without control light is coaxial with the larger one. When the control light is turned on, the smaller taper of about 1500μm long is driven to bend by optical force. The different poses of the smaller taper are shown in (b)–(h). The displacement of the tip of the smaller taper is measured and calibrated by a standard bare single-mode fiber with 125μm diameter. The result is showed in Fig. 4(i). We find that the process of bending can be divided into three phases. In the first phase from 0ms to 67ms, the tip is driven by light to move at a relative low average speed of about 40μm/s. This is because in this phase, the tip of the smaller taper suffers a stronger resistance, Van Der Waals force from the tip of the larger taper due to the nearness of two tips. In the second phase from 60ms to 330ms, the smaller tip is already far from the large one, and the Van Der Waals force tends to zero, and therefore the speed of the tip of the smaller taper is accelerated to a higher speed of about 202μm/s. At the same time, a new equilibrium, between the light forces and the resistances from both gravity and stress due to bend, is achieved. Thus the tip moves at an almost constant speed of about 202μm/s (see the middle of Fig. 4(i)). However, in the third phase, it is different from the case in the second phase that the moment of gravity and the stress due to bend become larger and larger, which resist the movement of the tip and therefore the speed of the tip decreases gradually and finally tends to zero.
Figure 5 shows the wave shape of signal captured by oscilloscope when the control light is turned off suddenly and the AOS enters the “on” state. We find from Fig. 5 that the turnon time of the AOS is about 760ms; the power of 650nm light passing through the AOS increases faster at beginning, but slower when AOS is close to full “on” state. Figure 6 shows the recovery process of the smaller fiber taper, where images (a)–(l) are cut from the same video mentioned above and (a) is a stable pose of the tip with 32mW control light coupled in; (b)–(l) are the poses of the tip after the control light is turned off. Figure 6(m) shows the displacement of the tip measured from (a)–(l). Similar to the bent process, the recovery process of the smaller taper is composed of three phases. The first phase when control light is just turned off: the tip is accelerated to recover at relative high speed (about 186μm/s). The second phase from 66ms to 333ms: the recovering speed of the tip decreases to about 29μm/s. The third phase from 530ms to 700ms: the tip is accelerated to recovery again at a speed of about 80μm/s, and finally, stops nearest to the larger taper. The physical actions in these three phases are similar to those described above for Fig. 4(i), but appear with a reverse sequence. It is observed that the turn on and turn off processes of the AOS can be repeated. (The detail can be viewed in the video).
The “critical” power, under which the signal light can be switched reliably and does not interfere with the switching, can be measured by such a way: increasing the power of control light slowly from zero and monitoring the signal; stopping the tuning of the control power at once when the signal begins to decrease; measuring the total power (including the control power and the signal one) coupled into the fiber. The measured total power is approximately the “critical” one. In our case, the “critical” power is about 7.50mW.
In conclusion, we have demonstrated a very simple all-optical switch (AOS) based on the bend effect of a nm fiber taper driven by outgoing light. The turnoff and recovery process of the AOS are studied. A turnoff time of about 500ms and a recovery time of about 760ms are observed in a prototype, which operates with a control power of 32mW at 980nm. The turn -off and -on time can be further shorten by sealing the AOS in a vacuum chamber, which reduces the resistance from air. Since this kind of AOS is all made by standard fiber, it is very convenient to directly connect with other fiber devices, which requires an ultra-low cost. Although the prototype of AOS demonstrated in our experiment is about 1500μm long, it is very easy to make the taper shorter and smaller, therefore reduce the size and the control power of AOS. In a word, such AOS has several intrinsic advantages: (1) very easy to make; (2) ultra-low cost; (3) very convenient to directly connect with other fiber devices; (4) low operation power; (5) relative small size. This new AOS could be used as an optical “relay” or a restorable optical “fuse”. But the stability of the AOS, such as the sensitivity to the external vibration and the degradation due to frequent switching needs to be further investigated.
Authors thank Prof. Zhigang Cai, Prof. Guozhong Lin, Fujuan Wang, Jiahui Wang for their assistance in the experiment and L. Tong for imparting flame-heated taper-drawing technique.
References and links
1. L. Tong, R. R. Gattass, J. B Ashcom, S. He, J. Lou, M. Shen, I. Maxwell, and E. Mazur, “Subwavelength-diameter silica wires for low-loss optical wave guiding,” Nature 426, 816–819 (2003). [CrossRef] [PubMed]
6. X. Jiang, L. Tong, G. Vienne, X. Guo, A. Tsao, Q. Yang, and D. Yang, “Demonstration of optical microfiber knot resonators,” Appl. Phys. Lett. 88, 223501 (2006). [CrossRef]
7. M. Sumetsky, Y. Dulashko, J. M. Fini, and A. Hale, “Optical microfiber loop resonator,” Appl. Phys. Lett. 86, 161108 (2005). [CrossRef]
10. X. Jiang, Q. Yang, G. Vienne, Y. Li, L. Tong, J. Zhang, and L. Hu, “Demonstration of microfiber knot laser,” Appl. Phys. Lett. 89, 143513 (2006). [CrossRef]
11. X. Jiang, Q. Song, L. Xu, J. Fu, and L. Tong, “Microfiber knot dye laser based on the evanescent-wave-coupled gain,” Appl. Phys. Lett. 90, 233501 (2007) [CrossRef]
16. M. H. Gibbs, Optical bistability: controlling light with light (Academic Press, 1985).
17. R. W. Boyd, Nonlinear Optics (Academic Press, 2002).
18. P. Bathelemy, M. Ghulinyan, Z. Gaburro, C. Toninelli, L. Pavesi, and D. S. Wiersma, “Optical switching by capillary condensation,” Nature 1, 172–175 (2007).
19. J. Liu, Q. Dai, Z. Meng, X. Huang, L. Wu, Q. Guo, W. Hu, S. Lan, A. V. Gopal, and V. A. Trofimov, “All-optical switching using controlled formation of large volume three-dimensional optical matter,” Appl. Phys. Lett. 92, 233108 (2008). [CrossRef]
21. M. D. Lukin, “Colloquium: Trapping and manipulating photon states in atomic ensembles,” Rev. Mod. Phys. 75, 457–472 (2003). [CrossRef]