Open Access
Add to BibSonomy
Abstract
We propose and experimentally demonstrate an all-optical switch based on a graphene-coated fiber Mach-Zehnder interferometer, where the phase of the signal light in one arm of the interferometer is changed by the heat generated from external pump light absorption by the graphene coating. The external pumping scheme allows efficient pump absorption with multiple layers of graphene coated on an ordinary fiber or a slightly tapered fiber without introducing significant additional signal loss. Without using any wavelength multiplexer/demultiplexer, the switch can be pumped at any convenient wavelength or even with broadband light. Our experimental device, which is based on a standard 125-μm-diameter single-mode fiber with a 5-mm-long graphene coating, can be switched with a pump power of 5.3 mW at an extinction ratio of 19 dB with no additional signal loss. The switching power is insensitive to the graphene coating’s length and can be reduced to 4.8 mW, with the fiber tapered to 40 μm. The measured switching powers agree well with the theoretical values obtained by treating the graphene coating as a uniform sheet of heat source without thickness. The switch’s response time decreases with the fiber diameter and inversely with the graphene coating’s length. The switch’s rise and fall times, based on a 40-μm tapered fiber with a 20-mm-long graphene coating, are 30 ms and 50 ms, respectively.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
As light-controlling-light devices, all-optical switches could find interesting applications in optical signal processing and optical communication networks. Fiber-based all-optical switches, in particular, have attracted much attention because of their compatibility to fiber transmission systems, simple implementation, and low cost. Nonlinear optical effects in silica fibers, such as cross-phase modulation [1,2] and the Kerr effect [3], have been explored for the realization of all-optical switches. These switches can operate at high speeds, but generally require very high pump powers (of the order of Watts and above). Highly nonlinear fibers, such as those doped with rare earth or transition metal ions, have been employed to form all-optical switches with pump powers in the range from milliwatts to hundreds of milliwatts [4,5].
The recent advances in the graphene technology offer new opportunities for the realization of all-optical devices. Being a 2D material, graphene possesses many appealing photonic properties, such as broadband absorption, tunable Fermi level, and large nonlinearity [6,7]. As graphene can be easily deposited or transferred onto a fiber, a graphene-incorporated fiber is an effective platform for the implementation of all-optical control devices. Over the last few years, a number of graphene-based all-optical modulators/switches with different performances have been reported. A graphene-coated microfiber modulator based on the Kerr effect of graphene [8] can offer 3-dB modulation with a pump power of 1.2 W at a speed of nanoseconds [9]. Modulators in the form of graphene-coated microfibers [10–12] and D-shaped fibers [13,14] based on the Pauli blocking principle of graphene [15] can provide ultrafast switching of the order of picoseconds, but the pump powers required are higher. For example, the microfiber modulator in [10] requires a pump power of 25 W for 2-dB modulation and the D-shaped fiber modulator in [13] requires a pump power of 720 W for 9-dB modulation. All-optical switching based on the photothermal effect of graphene has also been demonstrated, where the heat generated from the absorption of pump light by graphene serves to modulate the refractive index of a graphene-incorporated phase-sensitive fiber structure [16–18]. This mechanism can substantially reduce the switching power, but at the expense of the switching time. This type of all-optical switches suits applications that do not require fast response. With this mechanism, all-optical switching with a power of 11 mW, a response time of 9.1 ms, and a signal loss of 5.4 dB has been demonstrated with a Mach-Zehnder Interferometer (MZI) that contains a 10-μm-diameter microfiber covered with five layers of 5-mm-long graphene [16]. By using a ring resonator formed with a graphene-covered 5-μm-diameter microfiber, the response time is reduced to 0.3 ms, but the switching power required is 18 mW and the signal loss is as high as 6.78 dB/mm [17]. All-optical switching with a power of 4 mW, a response time of 20.7 ms, and a signal loss of 2.1 dB has also been demonstrated with a 16-μm-diameter microfiber Bragg grating covered with a 7.6-mm-long graphene coating [18]. The use of a graphene-attached microfiber together with a co-propagating pump in these studies [16–18] inevitably introduces a significant loss to the signal, as the signal is also absorbed by the graphene attached to the microfiber (in addition to the loss introduced by the microfiber itself). Furthermore, with in-fiber pumping, it is necessary to use a wavelength multiplexer/demultiplexer to combine/separate the signal and the pump at the two ends of the device, which adds complexity and cost to the device and, more importantly, restricts the pump wavelength that can be used. To address the loss issue, tungsten disulfide (WS2), which has strong absorption at 980-nm pump light and weak absorption near 1550-nm signal light, has been proposed to replace graphene [19]. The experimental device based on a WS2-deposited microfiber, however, shows a high signal loss of 3.7 dB and the switching power required is 57 mW [19].
In this paper, we propose an externally pumped all-optical switch with a fiber MZI based on the photothermal effect of graphene, where multiple layers of graphene are coated on one arm of the MZI and pump light is incident upon the graphene coating through free space. With this external pumping scheme, there is no need to use a microfiber or a D-shaped fiber to enhance graphene-light interaction. The graphene coating can be applied to an ordinary fiber or a slightly tapered fiber without introducing a significant loss to the signal. Without using a microfiber, the device is more robust and its performance is more stable. Furthermore, by doing away with the wavelength multiplexer/demultiplexer at the two ends of the device, the complexity and the cost of the device are significantly reduced and, more importantly, practically any light sources, including visible-light and white-light sources, can be used to pump the device. Our switch makes good use of the large light absorption rate and the high thermal conductivity of graphene for the achievement of low-power switching. The light absorption of graphene is ~2.3% for an atomic thickness of 0.34 nm [20]. In our study, we use ~20 layers of graphene, which amounts to a coating thickness of only ~7 nm and can absorb ~80% of pump light. The heat conversion from light absorption by graphene is highly efficient. The high thermal conductivity of graphene allows the heat generated locally to be distributed along the entire coating quickly, which provides a high utility of heat for our MZI. These characteristics of graphene allow the graphene coating to be modeled as a uniform sheet of heat source without thickness, which offers the highest possible heating efficiency for a light-absorbing coating. We analyze the effects of the diameter of the graphene-coated fiber and the length of the graphene coating on the switching power and perform experiments to verify the theoretical results. By using an ordinary 125-μm-diameter fiber coated with 5-mm-long graphene, we achieve a switching power of 5.3 mW at an extinction ratio of 19 dB with negligible graphene-induced signal loss. The switching power is insensitive to the length of the graphene coating and decreases with the fiber diameter. With the graphene-coated fiber tapered to a diameter of 40 μm, we reduce the switching power to 4.8 mW. The experimental results agree well with the theoretical predications. The use of a tapered fiber and a longer graphene coating can shorten the response time of the device from hundreds of milliseconds to tens of milliseconds. For example, the rise and the fall time of the switch based on a 40-μm tapered fiber with a 20-mm-long graphene coating are 30 ms and 50 ms, respectively. Our proposed external pumping scheme together with the excellent chemical stability and mechanical durability of graphene provides a practical solution for the realization of low-power all-optical switches. The unique features of the switch, in particular, pump-wavelength insensitivity and low signal loss, should facilitate the development and the applications of this type of switches.
2. Operation principle and analysis
Figure 1 is a schematic diagram of the structure of the proposed all-optical switch, which is an MZI formed with single-mode fibers. Signal light is launched into the two arms of the MZI equally with a 50/50 fiber coupler and collected at the output end with another 50/50 fiber coupler. A section of the fiber along one arm of the MZI (with the jacket removed) is wrapped with multiple layers of graphene and exposed to pump light. The heat generated from the absorption of the pump light by the graphene coating increases the temperature of the fiber, which causes a change in the phase of the signal light. Optical switching is achieved by producing a phase change of π at the signal wavelength.
We estimate the temperature distribution of the graphene-coated fiber section subject to the absorption of the pump light by using the heat transfer module in the commercial software COMSOL. The model we set up for the analysis is illustrated in Fig. 2, where the pump-absorbed graphene coating is treated as a heat source without thickness wrapped around the fiber. As graphene is an excellent heat conductor [21], the location and the size of the pump beam on the graphene coating are ignored, i.e., the heat generated along the graphene coating is assumed to be uniform. This model offers the highest possible heating efficiency for a light-absorbing coating.
Fig. 2 Model used for the analysis of the temperature distribution in a graphene-coated fiber, where the pump power absorbed by graphene is modelled as a uniform heat source without thickness wrapped around the fiber.
The cross-sectional temperature distribution T(r) is found by solving the following steady-state heat conduction equation:
where Q is the heat power density, i.e., the total pump-absorbed power divided by the area of the fiber surface, which is inversely proportional to the graphene length L, and ks = 1.38 W/(m·K) is the thermal conductivity of silica. The model ignores heat conduction in air, while takes into account air convection by enforcing the following boundary condition:where T0 is the temperature on the fiber surface (a variable), T∞ is the room temperature (set at 20 °C), and h is the air convection coefficient, whose value depends on the temperature difference T0 – T∞ and the fiber dimension and can be found from the data given in [22]. The values of T0 and h are continuously updated in the process of solving Eqs. (1) and (2). The above boundary condition means that, in the steady state, the conductive heat flux of the silica fiber equals to the convective heat flux of air on the fiber surface, i.e., at r = r0. To match the experimental conditions, we assume that 80% of the pump power is absorbed by graphene and converted into heat. Our simulation results show that the temperatures in the core and the cladding of the fiber are the same, which is consistent with the results in [16]. For a pump power of 5 mW, i.e., a heat power of 4 mW for 80% pump absorption, the calculated temperature change in a 125-μm-diameter fiber is ΔT = 17 °C for a 5-mm-long graphene coating. The fiber temperature increases linearly with the pump power.Given that the thermo-optic coefficient of a silica fiber is 1.1 × 10−5/°C at 1550 nm [16], the phase change Δϕ induced by the heat generated by the pump is
Figure 3(a) shows the calculated dependence of the phase shift on the pump power at two lengths of the graphene coating for an overall fiber diameter of 125 μm. As shown in Fig. 3(a), the pump power required for a π-phase change, i.e., the switching power, is ~4.5 mW, which is insensitive to the length of the graphene coating. According to Eq. (3), the phase shift is governed by ΔTL. With the air convection effect ignored, at a fixed pump power, the graphene length is inversely proportional to the temperature increase, so the phase shift incurred remains unchanged. The slightly larger phase shift obtained from a longer graphene coating, as shown in Fig. 3(a), is due to weaker air convection.Fig. 3 (a) Calculated phase shift as a function of pump power for two lengths of graphene coating on a 125-μm-diameter fiber; (b) calculated switching power as a function of the fiber diameter for a 10-mm-long graphene coating; and (c) calculated graphene-induced signal loss as a function of tapered fiber diameter for three lengths of graphene coating.
As shown in Fig. 3(a), milliwatt all-optical switching is possible with a graphene-coated 125-μm-diameter fiber. There should be room to further reduce the switching power and, at the same time, shorten the switching time by reducing the diameter of the graphene-coated fiber. We repeat the calculation for fibers with smaller diameters for a 10-mm-long graphene coating and present the results in Fig. 3(b). As shown in Fig. 3(b), by reducing the diameter of the coated fiber from 125 μm to 40 μm, the switching power is reduced from ~4.5 mW to ~4.1 mW. In practice, the fiber diameter can be reduced by tapering a standard fiber, but fiber tapering can introduce additional signal loss, mainly due to the absorption of signal by graphene. Figure 3(c) shows the dependence of the graphene-induced signal loss (at 1550 nm) on the tapered fiber diameter for different graphene lengths, where 20 layers of graphene wrapped around the fiber are assumed. To obtain the graphene-induced loss, we calculate the complex effective index of the fundamental mode of the fiber with the mode solver in COMSOL, where the graphene coating is treated as an interface with a pure conductivity [23]. The conductivity of a coating of N graphene layers is given by N × σmono [24], where σmono = 6.083 × 10−5 + j8.652 × 10−6 S is the conductivity of monolayer graphene at the wavelength 1550 nm [25]. The imaginary part of the complex effective index gives the loss [23]. As shown in Fig. 3(c), the graphene-induced signal loss is negligible at a fiber diameter larger than ~60 μm and increases rapidly as the fiber is tapered further. To avoid excessive graphene-induced signal loss, the tapered diameter should not be smaller than ~40 μm. For a 20-mm-long graphene coating on a 40-μm tapered fiber, the signal loss is ~0.7 dB.
3. Experimental setup
Figure 4 shows the experimental setup of the all-optical switch. The fiber MZI was formed by connecting two standard SMF-28 fibers to two 50/50 fiber couplers. The length difference between the two arms of the MZI was approximately 3 mm. Signal light was launched into the MZI via a 50/50 coupler and the output light from the other 50/50 coupler was measured. A section of the fiber along one arm of the MZI (with the jacket removed) was wrapped with multiple layers of graphene and clamped on two translation stages. The graphene-coated fiber was exposed to the output from a fiber-pigtailed pump source mounted on another translation stage. With this setting, it was easy to adjust the position of the pump beam in relation to the graphene coating to ensure that the entire pump beam is incident onto the graphene coating. A photodetector was placed under the graphene-coated fiber section to monitor the transmitted pump light. To determine the switching power, a broadband source (Amonics ALS-CL-15-B-FA) was used as the signal source and the output spectrum of the MZI was monitored with an optical spectrum analyzer (OSA) (Yokogawa AQ6370). To determine the response time of the switch, a tunable laser (HP 8168F) was used as the signal source and the pump laser was modulated by square waves generated from a signal generator (GW Instek AFG-2225). The output light from the MZI was detected with a photodetector and the modulated signals were displayed on an oscilloscope (Tektronix TBS 1102B).
The technique of wrapping graphene around a fiber is similar to that reported in [26]. A graphene monolayer sheet supported by a poly(methyl methacrylate) (PMMA) substrate (Hefei Vigon Material Technology Co.) was first released in deionized water. The fiber to be coated was then placed under the graphene sheet and pulled out of water together with the graphene sheet. After drying in air at room temperature for one hour and at 70 °C for 20 min, the graphene-covered fiber was submerged in acetone for 30 min to have the PMMA fully removed. Because of electrostatic attraction, the bare graphene sheet, which hanged onto the fiber like a flag, was naturally wrapped around the fiber, when it lost the support of the PMMA substrate. The length and the number of the graphene layers on the fiber were determined by the size of the starting graphene sheet. The pump source used was a pigtailed 1550-nm laser diode (Apico, CMP-LD-1550-BTF) or a 974-nm laser diode (Max-Ray Photonics, PFL-974-300-B). By comparing the pump powers passing through a bare fiber and the graphene-coated fiber, we were able to estimate the total absorption of the pump light and hence the number of the graphene layers. We found that the measured power loss due to scattering and reflection on a bare fiber was ~11%. The measured transmitted power through the graphene coating was ~9%, which indicates that ~80% of the pump light was absorbed by graphene, assuming the same scattering and reflection loss as a bare fiber. As a single graphene layer absorbs ~2.3% of the pump light, 80% absorption implies ~20 layers of graphene. The fiber section wrapped with graphene can be a bare SMF-28 fiber, which has an overall diameter of 125 μm, or a tapered SMF-28 fiber.
We should note that our method of wrapping graphene around a fiber does not produce a highly uniform graphene coating. Our switch, however, does not require an accurate control of the number of graphene layers and its uniformity along the coating. What matters is the amount of the pump light absorbed at the point of incidence, which is determined by the photodetector placed under the coating. Thanks to the high thermal conductivity of graphene, the heat generated by the pump at the point of incidence heats up the entire graphene coating quickly. The non-uniformity of the graphene coating does not affect much the global heating effect. The high tolerance on the quality of the graphene coating is another advantage of the external pumping scheme. For the same reason, the alignment of the pump beam is also relaxed. The exact location of the pump beam along the graphene coating is unimportant as long as the desired amount of the pump light is absorbed, which results in a robust and easy-to-operate switch.
4. Results and discussions
4.1 Switching power
We first present the results for a graphene-coated 125-μm fiber pumped by the 1550-nm laser diode. Figure 5(a) shows the output transmission spectra measured at several pump powers for a 5-mm-long graphene coating. As broadband light is used as the signal light, the transmission spectrum of the MZI consists of a number of interference fringes, which have a visibility of 19 dB. As shown in Fig. 5(a), the position of each interference fringe shifts by a half fringe (which corresponds to a phase change of π between the two arms of the MZI), as the pump power increases from 0 mW to 5.3 mW. Therefore, the switching power is 5.3 (with an uncertainty of ± 0.2 mW) and the extinction ratio of the switch is 19 dB. We conducted the experiments in our laboratory, where the temperature was 20 ± 0.4 °C. The small fluctuations in the ambient temperature should partly contribute to the uncertainties in the measurements of the locations of the fringes. We repeated the measurements for a 10-mm-long graphene coating and obtained almost the same results. The phase shift deduced from the fringe shift as a function of the pump power is shown in Fig. 5(b) for two lengths of the graphene coating. The measurement data shown in Fig. 5(b) agree reasonably well with the theoretical results shown in Fig. 3(a). The discrepancies should be mainly due to the use of an ideal 2D theoretical model in the analysis, which ignores the temperature variation along the graphene coating. Nevertheless, our experiments confirm that the dependence of the phase shift on the pump power is insensitive to the graphene length, as predicted by the theory. Our experimental results verify that the graphene coating can indeed be modeled approximately as a uniform sheet of heat source without thickness. We repeated the measurements with the 974-nm pump laser and obtained the same results within the measurement uncertainties, which confirms that the switch can operate equally well at different pump wavelengths.
Fig. 5 (a) Transmission spectra measured at several pump powers for a 125-μm fiber with a 5-mm-long graphene coating; and (b) phase shift as a function of the pump power measured for 125-μm fibers with 5-mm- and 10-mm-long graphene coatings, respectively.
We next present the results for graphene-coated tapered fibers pumped by the 1550-nm laser diode. We prepared a number of tapered fibers with the flame brushing technique, where the taper diameter of the fiber was varied by the fiber stretching length and speed. The tapered regions of the fibers were approximately 15 to 25 mm long. By controlling the width of the starting graphene sheet and monitoring the pump light passing through the graphene-coated fiber, we were able to wrap an enough number of graphene layers around the tapered fiber to ensure that at least 80% of the pump light was absorbed by the graphene coating. Figure 6(a) shows the switching powers measured for four chosen tapered fibers (with 100-μm, 80-μm, 60-μm, and 40-μm diameters, respectively) and the untapered fiber (with 125-μm diameter), each coated with 10-mm-long graphene, together with microscopic images of the graphene-coated sections of the fibers. As shown in Fig. 6(a), the switching power decreases from 5.3 mW to 4.8 mW (with an uncertainty of ± 0.2 mW), as the fiber diameter decreases from 125 μm to 40 μm. The amount of decrease in the switching power is 9.4%, which compares well with the theoretical value (8.8%), as shown in Fig. 3(b). Figure 6(b) shows the transmission spectra of the 40-μm fiber. The additional signal loss for the 40-μm fiber is 0.8 dB for a 10-mm-long graphene coating, which includes the absorption loss due to graphene and the taper loss.
Fig. 6 (a) Switching powers measured for five fibers with different diameters, each coated with 10-mm-long graphene, where the insets are microscopic images of the graphene-coated sections of the corresponding fibers; and (b) transmission spectra measured at different pump powers for the graphene-coated 40-μm fiber for a 10-mm graphene-coating length.
4.2 Response time
The response time of the switch depends on both the diameter of the graphene-coated fiber and the length of the graphene coating. Figure 7(a) shows the measured dependence of the rise and the fall time of the switch on the fiber diameter for a 10-mm-long graphene coating. By tapering the fiber diameter from 125 μm to 40 μm, the response time is shortened from hundreds of milliseconds to tens of milliseconds. We may further shorten the response by using a thinner fiber, such as a single-mode fiber designed for a shorter wavelength, which has a smaller core and should allow for stronger tapering without incurring a significant optical loss. Figure 7(b) shows the measured dependence of the rise and the fall time of the switch on the length of the graphene coating for a 125-μm fiber. At a fixed amount of heat generated by pump absorption, a longer graphene coating produces a lower temperature rise in the fiber (since the graphene coating acts as a uniform sheet of heat source, as discussed in Section 2 and confirmed by the experiments) and thus leads to a faster response. The response time can be shortened by ~50% by increasing the length of the graphene coating from 5 mm to 20 mm. Figure 7(c) shows an oscilloscope display of the signal waveforms together with the pump square waves for a 40-μm fiber with a 20-mm-long graphene coating. The rise and the fall time of this switch are 30 ms and 50 ms, respectively. The additional signal loss incurred by the graphene-coated fiber taper is 1.2 dB. The use of a special fiber for the coated section, such as a carbon nanotube fiber or a graphene fiber, to increase the thermal conductivity of the fiber [27,28] can also help to reduce the response time. There should also be room to increase the switching speed by devising an effective ventilation or heat dissipation system for the package of the switch.
Fig. 7 (a) Response times measured for five fibers with different diameters, each coated with 10-mm-long graphene; (b) response times measured for four 125-μm fibers with different graphene-coating lengths; and (c) oscilloscope display of the signal waveforms (black) together with the pump square waves (blue) for a 40-μm fiber with a 20-mm-long graphene coating.
5. Conclusion
We have proposed and demonstrated an all-optical switch based on a fiber MZI with one arm coated with multiple layers of graphene, where the phase of the signal light in the coated arm is changed by absorbing external pump light through the graphene coating. Unlike conventional in-fiber pumping that relies on the use of a graphene-coated microfiber, external pumping allows efficient pump absorption with an untapered or a slightly tapered fiber by increasing the number of graphene layers without incurring a large signal loss. External pumping does not need any wavelength multiplexer/demultiplexer and thus significantly reduce the complexity and the cost of the setup. More importantly, the operation of the switch does not depend on the wavelength of the pump light and practically any light source with a sufficient output power can be used as the pump source. Our analysis based on treating the graphene coating as a uniform sheet of heat source, as confirmed by experimental results, shows that the switching power is insensitive to the length of the graphene coating and can be lowered by reducing the diameter of the graphene-coated fiber. With ~80% pump power absorption (i.e., ~20 layers of graphene), the switching powers achieved with a standard 125-μm fiber and a 40-μm fiber are 5.3 mW and 4.8 mW, respectively, which are significantly lower than those of MZI switches based on graphene-covered microfibers. The response time of the switch can be shortened by reducing the fiber diameter and increasing the length of the graphene coating. The rise and the fall time of a switch based on a 40-μm fiber with a 20-mm graphene-coating length, which has an excess loss of 1.2 dB, are 30 ms and 50 ms respectively. It should be possible to further improve the performance of the switch by optimizing the parameters of the device and incorporating an effective heat dissipation scheme. With our proposed all-optical switching scheme, any 2D material with strong light absorption and excellent heat conductivity that can be wrapped around a fiber can be exploited for the development of all-optical control devices.
Funding
City University of Hong Kong (7005062); Research Grants Council, University Grants Committee, Hong Kong (CityU 11202014).
Acknowledgments
The authors thank Mr. Binghui Li for his valuable discussion and Mr. Wai Yip Lai and Dr. Wing Han Wong for their technical support.
References
1. J. E. Sharping, M. Fiorentino, P. Kumar, and R. S. Windeler, “All-optical switching based on cross-phase modulation in microstructure fiber,” IEEE Photonics Technol. Lett. 14(1), 77–79 (2002). [CrossRef]
2. A. Melloni, M. Chinello, and M. Martinelli, “All-optical switching in phase-shifted fiber Bragg grating,” IEEE Photonics Technol. Lett. 12(1), 42–44 (2000). [CrossRef]
3. B. J. Eggleton, R. E. Slusher, J. B. Judkins, J. B. Stark, and A. M. Vengsarkar, “All-optical switching in long-period fiber gratings,” Opt. Lett. 22(12), 883–885 (1997). [CrossRef] [PubMed]
4. M. J. F. Digonnet, R. W. Sadowski, H. J. Shaw, and R. H. Pantell, “Resonantly enhanced nonlinearity in doped fibers for low-power all-optical switching: a review,” Opt. Fiber Technol. 3(1), 44–64 (1997). [CrossRef]
5. Z. Zang and Y. Zhang, “Analysis of optical switching in a Yb3+-doped fiber Bragg grating by using self-phase modulation and cross-phase modulation,” Appl. Opt. 51(16), 3424–3430 (2012). [CrossRef] [PubMed]
6. K. S. Novoselov, V. I. Fal’ko, L. Colombo, P. R. Gellert, M. G. Schwab, and K. Kim, “A roadmap for graphene,” Nature 490(7419), 192–200 (2012). [CrossRef] [PubMed]
7. P. Avouris, “Graphene: electronic and photonic properties and devices,” Nano Lett. 10(11), 4285–4294 (2010). [CrossRef] [PubMed]
8. H. Zhang, S. Virally, Q. Bao, K. P. Loh, S. Massar, N. Godbout, and P. Kockaert, “Large nonlinear Kerr effect in graphene,” arXiv preprint arXiv:1203.5527 (2012).
9. S. Yu, X. Wu, K. Chen, B. Chen, X. Guo, D. Dai, L. Tong, W. Liu, and Y. Ron Shen, “All-optical graphene modulator based on optical Kerr phase shift,” Optica 3(5), 541–544 (2016). [CrossRef]
10. W. Li, B. Chen, C. Meng, W. Fang, Y. Xiao, X. Li, Z. Hu, Y. Xu, L. Tong, H. Wang, W. Liu, J. Bao, and Y. R. Shen, “Ultrafast all-optical graphene modulator,” Nano Lett. 14(2), 955–959 (2014). [CrossRef] [PubMed]
11. J.-H. Chen, B.-C. Zheng, G.-H. Shao, S.-J. Ge, F. Xu, and Y.-Q. Lu, “An all-optical modulator based on a stereo graphene–microfiber structure,” Light Sci. Appl. 4(12), e360 (2015). [CrossRef]
12. R. Wang, D. Li, M. Jiang, H. Wu, X. Xu, and Z. Ren, “All-optical intensity modulation based on graphene-coated microfibre waveguides,” Opt. Commun. 410, 604–608 (2018). [CrossRef]
13. H. Zhang, N. Healy, L. Shen, C. C. Huang, D. W. Hewak, and A. C. Peacock, “Enhanced all-optical modulation in a graphene-coated fibre with low insertion loss,” Sci. Rep. 6(1), 23512 (2016). [CrossRef] [PubMed]
14. C. Gao, L. Gao, T. Zhu, and G. Yin, “Incoherent optical modulation of graphene based on an in-line fiber Mach-Zehnder interferometer,” Opt. Lett. 42(9), 1708–1711 (2017). [CrossRef] [PubMed]
15. J. M. Dawlaty, S. Shivaraman, M. Chandrashekhar, F. Rana, and M. G. Spencer, “Measurement of ultrafast carrier dynamics in epitaxial graphene,” Appl. Phys. Lett. 92(4), 042116 (2008). [CrossRef]
16. X. Gan, C. Zhao, Y. Wang, D. Mao, L. Fang, L. Han, and J. Zhao, “Graphene-assisted all-fiber phase shifter and switching,” Optica 2(5), 468–471 (2015). [CrossRef]
17. Y. Wang, X. Gan, C. Zhao, L. Fang, D. Mao, Y. Xu, F. Zhang, T. Xi, L. Ren, and J. Zhao, “All-optical control of microfiber resonator by graphene’s photothermal effect,” Appl. Phys. Lett. 108(17), 171905 (2016). [CrossRef]
18. X. Gan, Y. Wang, F. Zhang, C. Zhao, B. Jiang, L. Fang, D. Li, H. Wu, Z. Ren, and J. Zhao, “Graphene-controlled fiber Bragg grating and enabled optical bistability,” Opt. Lett. 41(3), 603–606 (2016). [CrossRef] [PubMed]
19. K. Wu, C. Guo, H. Wang, X. Zhang, J. Wang, and J. Chen, “All-optical phase shifter and switch near 1550nm using tungsten disulfide (WS2) deposited tapered fiber,” Opt. Express 25(15), 17639–17649 (2017). [CrossRef] [PubMed]
20. F. Bonaccorso, Z. Sun, T. Hasan, and A. C. Ferrari, “Graphene photonics and optoelectronics,” Nat. Photonics 4(9), 611–622 (2010). [CrossRef]
21. A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, and C. N. Lau, “Superior thermal conductivity of single-layer graphene,” Nano Lett. 8(3), 902–907 (2008). [CrossRef] [PubMed]
22. M. K. Davis, M. J. F. Digonnet, and R. H. Pantell, “Thermal effects in doped fibers,” J. Lightwave Technol. 16(6), 1013–1023 (1998). [CrossRef]
23. Z. Chang and K. S. Chiang, “Experimental verification of optical models of graphene with multimode slab waveguides,” Opt. Lett. 41(9), 2129–2132 (2016). [CrossRef] [PubMed]
24. Y. Xiao, J. Zhang, J. Yu, H. Dong, Y. Wei, Y. Luo, Y. Zhong, W. Qiu, J. Dong, H. Lu, H. Guan, J. Tang, W. Zhu, and Z. Chen, “Theoretical investigation of optical modulators based on graphene-coated side- polished fiber,” Opt. Express 26(11), 13759–13772 (2018). [CrossRef] [PubMed]
25. Z. Chang, W. Jin, and K. S. Chiang, “Graphene electrodes for lithium-niobate electro-optic devices,” Opt. Lett. 43(8), 1718–1721 (2018). [CrossRef] [PubMed]
26. T. Hao and K. S. Chiang, “Graphene-based ammonia-gas sensor using in-fiber Mach-Zehnder interferometer,” IEEE Photonics Technol. Lett. 29(23), 2035–2038 (2017). [CrossRef]
27. G. Xin, T. Yao, H. Sun, S. M. Scott, D. Shao, G. Wang, and J. Lian, “Highly thermally conductive and mechanically strong graphene fibers,” Science 349(6252), 1083–1087 (2015). [CrossRef] [PubMed]
28. E. Mayhew and V. Prakash, “Thermal conductivity of high performance carbon nanotube yarn-like fibers,” J. Appl. Phys. 115(17), 174306 (2014). [CrossRef]
