Open Access
Add to BibSonomy
Abstract
Fan-in/fan-out (FI/FO) device with low crosstalk is essential for weakly coupled short-reach optical interconnect based on multicore fibers (MCF), for which the laser-direct-writing (LDW) technique is one of the preferred fabrication schemes. In this paper, the influence of FI/FO crosstalk on short-reach intensity-modulation/direction-detection MCF optical interconnection is firstly evaluated, and the crosstalk related to different refractive-index profiles of waveguides and misalignment is analyzed for LDW-FI/FO devices. Then low-crosstalk compact LDW-FI/FO devices matching 8-core MCF are fabricated, adopting multiple-scan method for waveguides with a flat-top refractive-index profile and aberration correction method for precise alignment. Owing to the low crosstalk, 8×100-Gbps optical interconnection over 10-km MCF is experimentally demonstrated with only 0.5-dB penalty compared to 10-km G.652D single-mode fiber transmission. Simulation results indicate that the transmission reach can be further extended to over 40 km. The proposed prototype system with low crosstalk is promising for high-speed optical interconnection applications.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Owing to the unprecedented increases in various services such as 5G, internet of things (IoT), cloud computing, and machine learning, short-reach optical interconnection applications are experiencing a rapid growth, for which intensity-modulation/direct-detection (IM/DD) transceivers over single-mode fibers (SMF) or multiple-mode fibers are always adopted to achieve simple system structure, low cost, and compact size [1]. Currently, typical 400G transmission schemes have been proposed based on the combination of multiple-level modulation format [2], wavelength division multiplexing (WDM) [2] or parallel single mode transmission (PSM) [3]. However, limited by achievable single-channel bit rate and available WDM or PSM channel numbers, continued bandwidth ascension beyond 400G is difficult along the road. Meanwhile, space-division-multiplexing (SDM) techniques based on few-mode fibers and multicore fibers (MCF) have been proposed as solutions to enhance the capacity, for which fiber cores or modes are explored as independent spatial channels [4]. Previous studies have shown that they can be promising candidates for optical interconnection applications. For examples, 10-km 100-Gbps/λ/core 7-core MCF transmission and 1.11-km 25-Gbps/λ/core 8-core MCF transmission have been experimentally demonstrated [5,6].
It should be noted that although high-speed long-haul transmission over strongly-coupled MCF has been proved to have great potential [7], the complex inter-core multiple-input-multiple-output (MIMO) digital-signal-processing (DSP) is not suitable for short-reach optical interconnections. Instead, weakly-coupled MCF transmission by significantly suppressing inter-core crosstalk is preferred because of high compatibility to conventional single-mode IM/DD transceivers, for which the MCF and matching fan-in/fan-out (FI/FO) devices are the key components. There may be three kinds of inter-core crosstalk, which are introduced by the MCF, FI/FO devices, and their connections. Previous studies have shown that increasing the core-to-core pitch or setting a low-refractive-index trench outside each core is effective to suppress the distributed crosstalk during MCF transmission [8]. As for the FI/FO devices, although multiple schemes have been proposed such as free-space optics [9] and fiber-bundle type [5], the fabrication method of laser-direct-writing (LDW) 3-dimensional (3D) waveguides may be the most promising scheme for the advantages of low insertion loss, compact size, high flexibility and repeatability. However, although LDW-FI/FO devices have been realized and their characteristics have been measured [10–12], the short-reach IM/DD transmission based on high-density MCF and low-crosstalk LDW-FI/FO devices has seldom been researched.
In this paper, the influence of FI/FO crosstalk on the performance of short-reach IM/DD MCF optical interconnection is firstly evaluated, and the crosstalk related to different refractive-index profiles of waveguides and misalignment for LDW-FI/FO device is analyzed. Then we successfully fabricate low-crosstalk compact LDW-FI/FO devices matching 8-core MCF adopting multiple-scan method for waveguides with flatted-top refractive-index profile and aberration correction method for precise alignment. Owing to the low crosstalk, 8×100-Gbps optical interconnection over 10-km MCF is experimentally demonstrated with only 0.5-dB penalty compared to 10-km G.652D SMF transmission. Simulation result indicates that the transmission reach can be further extended to over 40 km.
2. Analysis and design of LDW-FI/FO devices
In previous studies, the generation mechanisms of insertion loss for LDW-FI/FO devices have been widely discussed including waveguide scattering, waveguide-core misalignment, and mode field mismatching [10–12], but the influence of FI/FO crosstalk for high-density MCFs has seldom been discussed. In this section, we firstly analyze the influence of crosstalk on the short-reach IM/DD MCF optical interconnection, and then analyze the crosstalk related to the waveguide refractive-index profiles and misalignment for the LDW-FI/FO devices.
2.1 Influence of FI/FO crosstalk on IM/DD MCF optical interconnection
Simulations based on VPItransmissionMaker Optical Systems are carried out to evaluate the influence of FI/FO crosstalk on short-reach IM/DD MCF optical interconnection performance. The simulation setup and core arrangement of MCF are shown in Fig. 1(a) and 1(b), respectively. The parameters in the simulation are shown in Table 1. Eight transmitters (Tx) with 25.78 Gb/s on-off-keying (OOK) modulation at 1308.5 nm are used. The launching power is 1 dBm and the modulator extinction ratio is set to be 6 dB [13]. The loss of FI/FO devices is 2 dB/facet and the crosstalk for adjacent cores (CXTnearest) ranges from –40 to –20 dB. The high-density ring-core MCF is similar to that reported in [6]. The pitch is 31 µm with a cladding diameter of 125 µm, which is smaller than typical values of more than 40 µm [5,14]. The propagation attenuation is set to be 0.36 dB/km. The dispersion is 2.565 ps/nm/km at 1308.5 nm. The CXTnearest for MCF is set to be –65 dB/km. At the receiver (Rx), direct detection is applied and the back-to-back (BTB) receiving sensitivity is –20.35 dBm at Q2-factor limit of 11.8 dB.
Table 1. Parameters in the simulation.
Figure 2(a) shows the Q2-factor performance versus the crosstalk of FI/FO devices for BTB case. The receiving sensitivity penalties compared to that of BTB SMF transmission at Q2-factor limit of 11.8 dB are 0.08, 0.2, 0.56, and 2.76 dB when the CXTnearest is –40, –35, –30, and –25 dB, respectively. We can see that the FI/FO crosstalk may have significant influence on the performance of IM/DD MCF optical interconnection. When the CXTnearest is increased to be –20 dB, the signal cannot be received correctly. The Q2-factor performances for 10-km MCF optical interconnection are shown in Fig. 2(b). The receiving sensitivity penalties compared to that of BTB SMF transmission are 0.13, 0.36, 0.86, and 4.02 dB when the CXTnearest is –40, –35, –30, and –25 dB, respectively. Therefore, the crosstalk induced by the FI/FO devices may have greater influence on the system performance than that of distributed crosstalk and dispersion. It is necessary to analyze the generation mechanisms of FI/FO crosstalk and mitigate it as much as possible.
Fig. 2. (a) Q2-factor performance versus crosstalk of FI/FO devices for BTB case, and (b) Q2-factor performance versus crosstalk of FI/FO devices for 10-km MCF optical interconnection.
2.2 Analysis of low-crosstalk FI/FO for high-density MCF
The schematic structure of FI/FO device and its 3D layout are shown in Fig. 3 (a) and 3(b), which matches with the MCF core arrangement shown in Fig. 1(b). The waveguides of LDW-FI/FO device connect the fiber array (FA) and MCF cores from top to bottom in turn, which can avoid the waveguide coupling and crossing. The S-bends at top and bottom have relatively small lateral displacement, which can reduce the FI/FO length. Different aspects should be considered for the suppression of FI/FO crosstalk, including the S-bend route, the mode field, and the waveguide-core misalignment. When the bending radius of S-bend is smaller, the waveguides will be separated from each other in a shorter length and the waveguide coupling will be weaker [15]. Therefore, we should use a tight S-bend with a small radius. Here, the minimum bending radius is set to be 40 mm considering the bending loss. The mode field which is decided by the waveguide refractive-index profile and the waveguide-core misalignment which is the misalignment between the waveguides and MCF/FA cores will be analyzed in the following.
Fig. 3. (a) The schematic structure of FI/FO device. (b) 3D layout of the FI/FO chip. (c) The simulation model for waveguide coupling. (d) The simulation model for waveguide-core misalignment. (e) The graded refractive-index profiles with different gradients for the waveguides. (f) The waveguide crosstalk versus varying graded refractive-index profiles and core diameters. (g) The loss and crosstalk performance versus the waveguide-core misalignment for the waveguide with a graded refractive-index profile of α=7, Δ=0.36%, and R = 4 µm.
Simulations based on beam propagation method are used to analyze the crosstalk related to waveguide refractive-index profile and waveguide-core misalignment, the models for which are shown in Fig. 3(c) and 3(d), respectively. In previous studies, the refractive-index profile of the LDW-FI/FO waveguides are usually graded because of graded intensity distribution of the focus and heat accumulation effect [16,17]. We perform simulations for the waveguide coupling versus different graded refractive-index profiles. The graded refractive-index profile is assumed to be n(r)=ncore*[1-2Δ(r/R)α]1/2, r < R, where ncore is the maximum refractive index of the waveguide, Δ is the relative refractive index difference of core and cladding, R is the waveguide radius, and α is a variable to describe the gradient of the refractive-index profile. Figure 3(e) shows the graded refractive-index profiles with different gradients by setting the value of α from 1 to 10. Here Δ is 0.36%, ncore is 1.4518, and R is normalized to be 1. The crosstalk for two graded refractive-index waveguides with 5-mm coupling length is shown in Fig. 3(f). We can see that the coupling is weaker when the value of α is larger and the optimal refractive-index profile tends to be stepped. The simulation model shown in Fig. 3(d) is adopted to analyze the crosstalk versus different waveguide-core alignments. Figure 3(g) shows the loss and crosstalk when the waveguide-core misalignment ranges from 0 to 5 µm using waveguides with a graded refractive-index profile of α=7, Δ=0.36%, and R = 4 µm. We can see that the loss is mainly decided by the misalignment and the crosstalk is lower than –60 dB even the misalignment is 5 µm for the MCF. Therefore, the refractive-index profile of waveguides may be the major cause for the FI/FO crosstalk.
3. Fabrication and measuring of LDW-FI/FO devices
In order to realize low-crosstalk FI/FO devices for IM/DD MCF optical interconnection, we adopt multiple-scan method [10] to form the waveguides with stepped refractive-index profiles utilizing a femtosecond LDW system (InnoFocus Intelligent Laser Nano-Fabrication System), as shown in Fig. 4(a). In the fabrication process, 2.2-µJ/pulse femtosecond pulses with a pulse duration of 400 fs and repetition rate of 100 kHz from a femtosecond laser at 1030 nm are focused inside the boro-aluminosilicate glass chip (Corning Eagle XG) using a 0.4-NA dry objective with a speed of 12 mm/min. The fabrication depth is from about 120 to 200 µm to reduce the additional loss induced by the spherical aberration. The designed core diameter is 7.8 µm with about 50 scan times. The refractive-index profile is measured as shown in Fig. 4(b) [18] and the calculated mode field is inserted in Fig. 4(b). The waveguide refractive-index profile is reconstructed by imaging the multi-directional diffractive light field based on a galvanometric-scanner-assisted optical diffractive tomography (ODT) system. The mode area is about 62.7 µm2, which is slightly bigger than that of MCF measured to be from 53.2 to 55.6 µm2. We can see that the waveguide refractive-index profile presents to be flatted-top, which can effectively suppress the FI/FO crosstalk according to the analysis in section 2.2.
Fig. 4. (a) LDW system. (b) The measured waveguide refractive-index profile and calculated mode field by multiple-scan method.
Since we use a dry objective, the aberration of the laser focus should be corrected to guarantee the waveguide-core alignment, as shown in Fig. 5(a) [19]. The focus depth is changed because of the laser light refraction at the interface between air and glass. The real focus depth can be given by D ≈ (n2/n1)F, where n2 and n1 are the refractive indexes of the glass chip and air, respectively, and F is the virtual focus depth. Before aberration correction, the waveguide depth is bigger than we want, whereas the waveguide location at the horizontal direction does not change, so the cross-section of FI/FO looks like ellipse, as shown in Fig. 5(b). After aberration correction when the coefficient n2/n1 is 1.47, the measured distance between core #2 and #7 is 74.7 µm and the measured distance between core #4 and #5 is 31.1 µm, which match the designed values in Fig. 1(b). Simulations based on beam propagation method are utilized to show the tolerance of n2/n1 in Fig. 6(a). We can see that the loss induced by the core location deviation is below 0.5 dB when the value of n2/n1 ranges from 1.414 to 1.531.
Fig. 5. (a) Light path diagram of aberration for the LDW system, (b) cross-sections of FI/FO at the 8-core MCF side before and after aberration correction, respectively.
Fig. 6. (a) The loss induced by the core location deviation versus the value of n2/n1. (b) The measured transfer matrix for a pair of FI/FO devices.
The pitch of FA is 127 µm and the size of FI/FO chip is 10.2×0.9 mm2. The 8-core MCF and FA are coupled to the end-faces of the FI/FO chip and fixed using refractive-index-matched UV adhesives. It should be noted that no lens is used when coupling the chips to the fibers. The measured insertion loss for each packaged FI/FO device at 1310 nm ranges from 1.9 to 2.6 dB, which mainly induced by waveguide scattering and the mode field mismatching and misalignment between the waveguides and MCF/FA. The measured waveguide propagation loss is about 0.4 dB/cm by subtracting the loss of 1-cm waveguide from 2-cm waveguide. The insertion loss is from 1 to 1.3 dB under perfect waveguide-core alignment conditions, which means that about 0.6 to 0.9-dB loss is induced by mode-field mismatching. And about 0.6 to 1.6-dB loss is mainly caused by the waveguide-core misalignment. The insertion loss is slightly bigger than the typical value of 1.2 dB with single-scan method [12], which is mainly due to waveguide-core mismatching and the relative displacement caused from the package process. Optimizing multiple-scan arrangement and using advanced package process can reduce the insertion loss. The transfer matrix for directly connecting the FI to FO devices is shown in Fig. 6(b). We can see that the crosstalk for a pair of FI/FO devices is lower than –31 dB, owing to the flatted-top refractive-index profile for the waveguides.
4. 8×100-Gbps MCF optical interconnection architecture and transmission performance
We establish the prototype of 8×100-Gbps optical interconnection based on the fabricated MCF and LDW-FI/FO devices, which is similar to the setup in Fig. 1(b). Figure 7(a) is the picture of the LDW-FI/FO devices and Fig. 7(b) is the overall picture of the 10-km MCF and FI/FO devices. Figure 7(c) shows the picture of the cross-section of the fabricated MCF measured by fiber end-face detector (SGX-7000C). Figure 7(d) shows the measured refractive-index distribution for the cross-section and Fig. 7(e) shows the refractive-index profile for one core (IFA-100). Eight 100GBASE-ER4 lite transceivers (ZTE SM-LWDM-100GE-C) are used for the Txes and Rxes. Each Tx can generate 4-wavelength-channel On-Off Keying (OOK) single-mode optical signals modulated by 25.78125-GBaud pseudo-random binary sequence (PRBS, 231-1) at 1296.2, 1299.5, 1305.2, and 1308.5 nm. As for the MCF, the measured propagation attenuation of each core at 1310 nm ranges from 0.34 to 0.36 dB/km. The dispersion is ranging from 1.4086 to 2.5652 ps/nm/km when the wavelength is from 1296.2 to 1308.5 nm for a typical core. The CXTnearest is evaluated to be –65 dB by CXTnearest ≈ λκ2RL/(πneffd) [20], where λ is the wavelength of 1310 nm, neff is the effective refractive index of the guiding mode with the value of 1.4492, R is the bending radius with the value of 0.2 m in the fiber spool, L is set to 1 km, d is the pitch of 31 µm, and κ is the modal coupling coefficient [15]. The crosstalk of non-adjacent cores with lager distances can be negligible. After the transmission over 10-km 8-core MCF, the signals are demultiplexed by the FO device and detected by Rxes. The loss of each channel for the 10-km MCF optical interconnection is from 8.5 to 9.2 dB, including about 4.7 dB to 5.4-dB loss for the FI and FO devices, 3.4 to 3.6-dB loss for the MCF transmission, and 0 to 0.7-dB loss for the connections of FI/FO devices and MCF. No optical amplifiers are used. In order to measure the receiving sensitivity of the transmission system, a single-mode variable optical attenuator (EXFO FVA 600) is placed before the Rxes to adjust the received power.
Fig. 7. (a) Picture of the FI/FO device, (b) picture for both the FI/FO devices and 10-km 8-core MCF, (c) picture of the cross-section of the MCF, (d) measured refractive-index distribution for the cross-section of MCF, and (e) refractive-index profile of one core.
The spectrums are measured at the output of core #1 with optical spectrum analyzer (VIAVI MTS 8000E) for the output of Tx, BTB transmission, and 10-km MCF optical interconnection with all cores excited, as shown in Fig. 8(a). The eye diagrams are measured with the digital sampling oscilloscope (Tektronix DSA8300) for the three cases with only signal of 1296.2 nm excited, as shown in Fig. 8(b)–8(d). Then BER performance for BTB transmission is evaluated by directly connecting FI and FO devices. As a reference, BTB SMF transmission by directly connecting the Tx and Rx is also performed. The experimental results are shown in Fig. 9(a)–9(d). We can see that the receiving sensitivities at the BER limit of 5e-5 of BTB transmission are about 0.4 to 0.5 dB worse than those of BTB SMF transmission at all 4 wavelengths. Experiments for 10-km MCF and G.652D SMF transmissions are also conducted, as shown in Fig. 9(a)–9(d). Compared to the 10-km G.652D SMF transmission case, the penalties of 10-km MCF optical interconnection at BER of 5e-5 range from 0.4 to 0.5 dB for all 4 wavelengths, so the inter-core crosstalk and chromatic dispersion during the 10-km MCF optical interconnection only have slight influence on the overall system performance. And we can see that the experimental results are consistent with those in the simulation.
Fig. 8. (a) Spectrums, and eye diagrams at the output of core #1 for (b) Tx, (c) BTB, and (d) 10-km MCF optical interconnection.
Fig. 9. BER curves for BTB and 10-km MCF optical interconnection compared to BTB SMF and 10-km G.652D SMF transmissions at the wavelengths of (a) 1296.2 nm, (b) 1299.5 nm, (c) 1305.2 nm, and (d) 1308.5 nm.
Then, the potential transmission reach is investigated based on the measured FI/FO transfer matrix and the performance of commercial transceivers. Due to the difference from the total crosstalk and transmission loss, core #2 and #5 have the worst and best performance, respectively. The simulated Q2-factor performances for the two cores are shown in Fig. 10. We can see that the transmission reach limits for core #2 and #5 are 41.6 km and 44 km, respectively. And the corresponding received power for core #2 and #5 is –19.0 dBm and –19.5 dBm, respectively. It illustrates that the penalties induced by dispersion and inter-core crosstalk are still small even when the transmission reach is over 40 km. Although the use of DSP would be acceptable for 40-km transmission reach, we still explore the potential maximum transmission reach for the proposed MCF optical interconnection to better evaluate its application in different transmission scenarios.
Fig. 10. Q2-factor performance versus transmission reach using the commercial transceiver, fabricated FI/FO devices and MCF.
5. Conclusion
In this paper, we have designed and fabricated compact and low-crosstalk LDW-FI/FO devices, based on which a real-time IM/DD 8×100-Gbps MCF optical interconnection is experimentally demonstrated. We firstly evaluate the influence of FI/FO crosstalk on the performance of short-reach IM/DD MCF optical interconnection and analyze the crosstalk related to different waveguide refractive-index profiles and misalignment for the LDW-FI/FO devices. Then we successfully fabricate low-crosstalk compact LDW-FI/FO devices matching 8-core MCF adopting multiple-scan method for waveguides with flatted-top refractive-index profile and aberration correction method for precise alignment. Owing to the low crosstalk, 8×100-Gbps optical interconnection over 10-km MCF is experimentally demonstrated with only 0.5-dB penalty compared to 10-km G.652D SMF transmission. Simulation result indicates that the transmission reach can be further extended to over 40 km. The fabricated FI/FO chip has a length of slightly larger than 10 mm, which need to be further reduced by utilizing high-refractive-index-difference tightly-bending waveguides and two-dimensional small-pitch Tx/Rx arrays to be integrated in commercial transceivers with package specifications such as Quad Small Form Factor Pluggable-Double Density (QSFP-DD) [21]. We believe that the proposed prototype system based on low-crosstalk compact LDW-FI/FO devices is promising for high-speed optical interconnection applications.
Funding
National Natural Science Foundation of China (61627814, 61690194, 61901009, 61771024, U20A20160); China Postdoctoral Science Foundation (2020M680236, BX20200003); Yangtze Optical Fibre and Cable Joint Stock Limited Company (SKLD2003).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon request.
References
1. X. Zhou, R. Urata, and H. Liu, “Beyond 1 Tb/s Intra-Data Center Interconnect Technology: IM-DD OR Coherent?” J. Lightwave Technol. 38(2), 475–484 (2020). [CrossRef]
2. “IEEE Standard for Ethernet,” IEEE Std 802.3cu-2021, Amendment 11, doi: 10.1109/IEEESTD.2021.9381783 (2021).
3. T. Shi, T. Su, N. Zhang, C. Hong, and D. Pan, “Silicon Photonics Platform for 400G Data Center Applications,” in Optical Fiber Communications Conference and Exposition (OFC)2018, paper M3F.4.
4. D. J. Richardson, J. M. Fini, and L. E. Nelson, “Space-division multiplexing in optical fibres,” Nat. Photonics 7(5), 354–362 (2013). [CrossRef]
5. R. Lin, J. V. Kerrebrouck, X. Pang, M. Verplaetse, O. Ozolins, A. Udalcovs, L. Zhang, L. Gan, M. Tang, S. Fu, R. Schatz, U. Westergren, S. Popov, D. Liu, W. Tong, T. D. Keulenaer, G. Torfs, J. Bauwelinck, X. Yin, and J. Chen, “Real-time 100 Gbps/λ/core NRZ and EDB IM/DD transmission over multicore fiber for intra-datacenter communication networks,” Opt. Express 26(8), 10519–10526 (2018). [CrossRef]
6. T. Hayashi, T. Nakanishi, K. Hirashima, O. Shimakawa, F. Sato, K. Koyama, A. Furuya, Y. Murakami, and T. Sasaki, “125-µm-Cladding Eight-Core Multi-Core Fiber Realizing Ultra-High-Density Cable Suitable for O-Band Short-Reach Optical Interconnects,” J. Lightwave Technol. 34(1), 85–92 (2016). [CrossRef]
7. G. Rademacher, R. S. Luís, B. J. Puttnam, R. Ryf, S. van der Heide, T. A. Eriksson, N. K. Fontaine, H. Chen, R. Essiambre, Y. Awaji, H. Furukawa, and N. Wada, “172 Tb/s C + L Band Transmission over 2040km Strongly Coupled 3-Core Fiber,” in Optical Fiber Communication Conference Postdeadline Papers 2020, paper Th4C.5.
8. K. Saitoh and S. Matsuo, “Multicore Fiber Technology,” J. Lightwave Technol. 34(1), 55–66 (2016). [CrossRef]
9. W. Klaus, J. Sakaguchi, B. J. Puttnam, Y. Awaji, N. Wada, T. Kobayashi, and M. Watanabe, “Free-space coupling optics for multicore fibers,” IEEE Photonics Technol. Lett. 24(21), 1902–1905 (2012). [CrossRef]
10. Y. Ling, S. Yuan, and S. J. Ben Yoo, “Low-Loss Three-Dimensional Fan-in/Fan-out Devices for Multi-Core Fiber Integration,” in Optical Fiber Communication Conference (OFC)2021, paper Th1A.37.
11. R. R. Thomson, H. T. Bookey, N. D. Psaila, A. Fender, S. Campbell, W. N. MacPherson, J. S. Barton, D. T. Reid, and A. K. Kar, “Ultrafast-laser inscription of a three-dimensional fan-out device for multicore fiber coupling applications,” Opt. Express 15(18), 11691–11697 (2007). [CrossRef]
12. A. Ross-Adams, S. Gross, B. J. Puttnam, R. S. Luís, G. Rademacher, and M. J. Withford, “Enabling Future Fiber Networks Using Integrated Ultrafast Laser-Written Multicore Fiber Fan-outs,” in 14th Pacific Rim Conference on Lasers and Electro-Optics (CLEO PR 2020), paper C3H_3.
13. “IEEE Standard for Ethernet,” IEEE Std 802.3-2018, Section 6, Table 88-7 (2018).
14. B.J. Puttnam, R.S. Luis, W. Klaus, J. Sakaguchi, J.-M.D. Mendinueta, Y. Awaji, N. Wada, Y. Tamura, T. Hayashi, M. Hirano, and J. Marciante, “2.15 Pb/s transmission using a 22 core homoge-neous single-mode multi-core fiber and wideband optical comb,” in European Conference on Optical Communication (ECOC) Postdeadline Papers 2015, paper PDP 3-1.
15. P. D. McIntyre and A. W. Snyder, “Power transfer between optical fibers,” J. Opt. Soc. Am. 63(12), 1518–1527 (1973). [CrossRef]
16. A. Ruiz de la Cruz, A. Ferrer, W. Gawelda, D. Puerto, M. Galván Sosa, J. Siegel, and J. Solis, “Independent control of beam astigmatism and ellipticity using a SLM for fs-laser waveguide writing,” Opt. Express 17(23), 20853–20859 (2009). [CrossRef]
17. S. M. Eaton, H. Zhang, M. L. Ng, J. Li, W. Chen, S. Ho, and P. R. Herman, “Transition from thermal diffusion to heat accumulation in high repetition rate femtosecond laser writing of buried optical waveguides,” Opt. Express 16(13), 9443–9458 (2008). [CrossRef]
18. D. Dong, X. Huang, L. Li, H. Mao, Y. Mo, G. Zhang, Z. Zhang, J. Shen, W. Liu, Z. Wu, G. Liu, Y. Liu, H. Yang, Q. Gong, K. Shi, and L. Chen, “Super-resolution fluorescence-assisted diffraction computational tomography reveals the three-dimensional landscape of the cellular organelle interactome,” Light: Sci. Appl. 9(1), 11 (2020). [CrossRef]
19. C. Hnatovsky, R. S. Taylor, E. Simova, V. R. Bhardwaj, D. M. Rayner, and P. B. Corkum, “High-resolution study of photoinduced modification in fused silica produced by a tightly focused femtosecond laser beam in the presence of aberrations,” J. Appl. Phys. 98(1), 013517 (2005). [CrossRef]
20. T. Hayashi, T. Sasaki, E. Sasaoka, K. Saitoh, and M. Koshiba, “Physical interpretation of intercore crosstalk in multicore fiber: Effects of macrobend, structure fluctuation, and microbend,” Opt. Express 21(5), 5401–5412 (2013). [CrossRef]
21. www.QSFP-DD.com
