96-Channel on-chip reconfigurable optical add-drop multiplexer for multidimensional multiplexing systems
-
Weike Zhao
, Shi Zhao
, Xiaolin Yi
, Guangze Wu
, Yaocheng Shi
and Daoxin Dai
Abstract
The multi-dimensional multiplexing technology is very promising for further increasing the link capacity of optical interconnects. A 96-channel silicon-based on-chip reconfigurable optical add-drop multiplexer (ROADM) is proposed and demonstrated for the first time to satisfy the demands in hybrid mode/polarization/wavelengthdivision-multiplexing systems. The present ROADM consists of a six-channel mode/polarization de-multiplexer, a 6 × 16 array of microring-resonator (MRR)-based wavelength-selective switches, and a six-channel mode/polarization multiplexer. With such a ROADM, one can add/drop optical signals to/from any channels of the multimode bus waveguide arbitrarily. For the designed and fabricated ROADM chip, there are more than 1000 elements integrated monolithically, including 96 MRRs, 576 waveguide crossings, 192 grating couplers, 96 micro-heaters, 112 pads, six polarization-splitter-rotators (PSRs), four asymmetric adiabatic couplers and four asymmetric directional couplers. For any channel added/dropped with the fabricated ROADM, the on-chip excess loss is about 5–20 dB, the inter-mode crosstalk is <−12 dB, and the inter-wavelength crosstalk is <−24 dB. The system experiments are demonstrated by using 10-GBaud quadrature phase shift keying (QPSK) signals, showing that the observed optical signal noise ratio (OSNR) power penalties induced by the ROADM are less than 2 dB at a BER of 3.8 × 10−3.
1 Introduction
The desire for high-capacity communication pushes the development of optical interconnects technology. Due to the Shannon limit and the erbium-doped fiber amplifier (EDFA) bandwidth limit, the widely used wavelength-division-multiplexing (WDM) technology hits its bottleneck [1]. The emerging mode-division-multiplexing (MDM) technology, which utilizes multiple modes as independent signal channels, provides a new dimension to further increase the traffic capacity [2–4]. Nowadays, the development of MDM has made considerable progress. Many key components have been developed for MDM systems, such as mode (DE)multiplexers (MUXs) [5, 6], multimode waveguide bends [7, 8], mode converters [9, 10], multimode waveguide crossings [11–13], mode filters [14], multimode power splitters [15, 16], etc. The system-level verification also proves the potential of the MDM systems [17, 18]. Currently, it is becoming more and more viable and attractive to develop a multi-dimensional multiplexing technology by combining m mode-channels, dual polarization-channels as well as n wavelength-channels together [19–21], because it increases the channel number greatly from n to 2 mn.
Consistent with its counterpart in WDM systems, there are some key devices/modules in hybrid MDM-WDM-PDM systems, including hybrid MUXs [22–25], hybrid switches [26–28], hybrid modulators [29], etc., which are usually constructed on the combination of high-performance mode/polarization/wavelength-manipulated elements, including mode MUX/DEMUXs [30–33], polarization rotators (PRs) [34–36], polarization beam splitters (PBSs) [37, 38], polarization splitter-rotators (PSRs) [39–41], micro-ring resonators (MRRs) [42, 43], arrayed-waveguide gratings (AWGs) [44–46], etc. Reconfigurable optical add-drop multiplexer (ROADM), which enables any channel to be switched and routed flexibly, is a key element in multiplexed optical communication systems [47–52]. In particular, recently ROADM for MDM systems has also attracted intensive attention [53–55]. In [53], a four-channel ROADM for MDM systems was demonstrated for the first time by integrating a pair of mode (DE)MUXs and four Mach–Zehnder switches (MZSs). Later, in [54] a ROADM for hybrid WDM-MDM systems was demonstrated for the first time by integrating a pair of mode (DE)MUXs and four microring-resonator-based wavelength-selective switches. However, it is still even challenging to develop a ROADM chip for multidimensional multiplexing systems because many wavelength-channels, a number of mode-channels, and dual polarization-channels are involved.
In this paper, we propose and demonstrate a 96-channel silicon-based on-chip ROADM for the first time to satisfy the demands in hybrid MDM-WDM-PDM systems. Here three modes, dual polarizations, and 16 wavelengths are involved. Accordingly, the present ROADM integrates a pair of six-channel hybrid MDM-PDM (DE)MUXs, and a 6 × 16 array of MRR-based wavelength-selective switches. In particular, the six-channel hybrid MDM-PDM (DE)MUXs are realized by introducing asymmetric adiabatic couplers for TE-polarization modes and asymmetric directional couplers (ADCs) for TM-polarization modes, regarding that there might be some notable polarization mode conversion due to the non-vertical sidewalls of the fabricated silicon photonic waveguides. For the designed and fabricated ROADM chip, there are more than 1000 elements integrated monolithically, including 96 MRRs, 576 waveguide crossings, 192 grating couplers, 96 micro-heaters, 112 pads, six polarization-splitter-rotators (PSRs), four asymmetric adiabatic couplers and four ADCs. With such a ROADM, one can add/drop optical signals to/from any channels of the multimode bus waveguide arbitrarily by switching the corresponding MRR-based optical switch. The fabricated ROADM chip shows on-chip excess losses (ELs) of 5–20 dB, inter-mode crosstalk of <−12 dB and inter-wavelength crosstalk of <−20 dB for adding any one of all 96 channels. The system experiments are demonstrated by using 10-GBaud quadrature phase shift keying (QPSK) signals, showing that the observed optical signal noise ratio (OSNR) power penalties induced by the ROADM are less than 2 dB at a BER of 3.8 × 10−3. The proposed scheme can be extended easily for even high capacity by adopting more mode- or wavelength-channels.
2 Principle and structural design
The present 96-channel ROADM for hybrid WDM-PDM-MDM systems is designed on the silicon-on-insulator (SOI) platform with a 220-nm-thick top-silicon layer and a 2 μm silica buffer layer. As shown in Figure 1, the ROADM consists of a pair of 6-channel hybrid MDM-PDM DEMUX, and a 6 × 16 array of MRR-based wavelength-selective switches. The 6-channel hybrid MDM-PDM DEMUX operates with three transverse electric (TE) modes (i.e., the TE0, TE1, and TE2 modes) and three transverse magnetic (TM) modes (i.e., the TM0, TM1, and TM2 modes). The 6 × 16 MRR-based switches array has six rows corresponding to the six mode/polarization channels, and each row contains 16 cascaded MRR-based switches. The resonant wavelengths of these 16 cascaded MRRs increase from λ 1 to λ 16 with a channel spacing of Δλ ch. Microheaters are used for the resonant wavelength tuning of each MRR-based switch, as shown in Figure 1, Pad G n (n = 1–16) is the shared ground electrode of the nth wavelength channels, and H m:n are the positive electrodes for mth mode and nth wavelength channels. Moreover, 6 × 96 waveguide crossings working on the principle of multimode interference are used for the crossing connection of the add/drop waveguides and the single-mode waveguides (SMWs) of the switches array [56]. For the hybrid MDM-PDM-WDM ROADM systems, data propagating along the multimode bus waveguide (MBW) are carried by 6 mode-, dual-polarization- and 16 wavelength-channels. When data arrives at the input MBW of the mode/polarization DEMUX, they are first demultiplexed into six TE0 modes by using the mode DEMUX together with the PRs/PSR, and then pass through the six SMWs of the 6 × 16 switches array, respectively. For each TE0 mode propagating through the SMW, the 16 wavelength-channels (λ 1, … λ 16) are dropped to port D m:n (n = 1, … 16) or go through the SMW directly by switching the resonant wavelength of the corresponding MRR to/away from the target λ n . After passing through the switches array, these six TE0 modes from the SMWs are then multiplexed into the MBW by another mode MUX combined with PRs/PSRs. Similarly, when the nth wavelength-channel (λ n ) is dropped to port D m−n , one can also add local signals carried by wavelength λ n from port A m:n to the same channel.

Configuration of the proposed ROADM for the hybrid MDM-PDM-WDM systems.
2.1 Mode (DE)multiplexer
Various types of high-performance modes (DE)MUX on silicon have been proposed previously [30–33]. For these proposed modes (DE)MUX, as it might be noticed, the silicon photonic waveguides used are assumed to have a vertical sidewall. Unfortunately, the fabricated silicon photonic waveguides usually have angled sidewalls due to the imperfect dry etching process [57]. As pointed out in [57], there might be some mode hybridness and mode conversion when light propagates in a tapered waveguide with angled sidewalls. Figure 2(a) and (b) show the calculated effective indices N eff of the guided modes in the 220-nm-thick silicon photonic waveguides with a sidewall angle of 90° and 86°, respectively. It can be seen that the TE1-TM0 and TE2-TM1 mode hybridness happens in the regions around the core width of 0.65 μm and 0.87 μm. When the waveguide is tapered around these two regions, the mode conversion happens, which might introduce some intermode crosstalk. This introduces lots of trouble for the design for the mode (DE)MUXes for which higher-order modes are involved. In particular, the TM0-TE1 mode conversions in a silicon photonic waveguide linearly tapered from 0.6 to 0.7 μm are analyzed, as shown in Figure 2(c). Meanwhile, Figure 2(d) shows the TM1-TE2 mode conversion in a silicon photonic waveguide linearly tapered from 0.82 μm to 0.92 μm. Here different lengths (L t ) and sidewall angles (θ) are chosen. It shows that the mode conversion can be reduced by choosing nearly vertical sidewalls and a short taper. The crosstalk is less than −22 dB for the TM0-TE1 mode conversion when the angle θ > 88° and the taper length L t < 6 μm. In contrast, when θ > 88° and L t < 5 μm, the crosstalk is less than −15 dB for the TM1-TE2 mode conversion.

Calculated mode effective index (N eff) for the silicon photonic waveguides with 90° sidewalls (a) and 86° sidewalls (b). Calculated efficiency of the TM0-TM0/TE1 in a waveguide linearly tapered from 0.6 μm to 0.7 μm (c) and TM1-TM1/TE2 mode conversion in a waveguide linearly tapered from 0.82 μm to 0.92 μm (d), here different lengths (L t ) and sidewall angles (θ) are chosen.
According to the analysis above, the adiabatic coupling region of (DE)MUX should be designed to exclude any mode hybridness region, while a short taper should be used when there is some mode hybridness for the waveguide whose width is varied from w 1 to w 2. Considering the mode effective index of TE modes are more sensitive to the waveguide width than that of TM modes, the mode/polarization (DE)MUXs are designed by combining the asymmetric adiabatic couplers for TE modes and ADCs for TM modes here. With such a hybrid scheme, one can achieve a mode MUX with a broad bandwidth and a large fabrication tolerance. Besides, one can also jump over the mode hybridization region with a nonadiabatic taper used in the uncoupled region to minimize the undesired intermode conversion. Figure 3(a) schematically shows the six-channel mode/polarization (DE)MUX, The TM2, TM1, TE2, and TE1 higher-order modes are decoupled to the TM0 and TE0 modes in four access waveguides successively, and the TM0 modes are further converted to the TE0 modes with the corresponding PRs. Moreover, the TE0 and TM0 modes that remained in the MBW are handled by a PSR, which converts the input TM0 mode to the TE0 mode at the output end [41]. Figure 3(b) shows the schematic diagram of the ADC-based mode DEMUX used for the TM modes. Here the wide-core width w b and the narrow-core width w a are chosen according to the phase-matching condition [30], and the coupling region length is chosen optimally. Figure 3(c) schematically shows the asymmetric-adiabatic-coupler-based mode DEMUX for the TE modes. The wide core is tapered from w b1 to w b2 with a length of L, and the narrow core is tapered from w a1 to w a2 correspondingly [33], while the gap between the wide core and narrow core is set as w g. The parameters chosen for the four couplers are listed in Table 1, and a 6-μm-long nonadiabatic taper is used to vary the waveguide width from 0.62 to 0.93 μm. Figure 4 (a–d) show the simulated light propagation for the TM2, TM1, TE2, and TE1 modes in the designed couplers when operating at the wavelength of 1550 nm. Finally all the modes are efficiently coupled to the TM0 or TE0 mode in the access waveguide. The transmissions of these four modes are shown in Figure 5(a–d). It can be seen that the ELs for the TM2, TM1, TE2, and TE1 modes are less than 1.3 dB, 0.95 dB, 0.02 dB, and 0.06 dB while the crosstalks are less than −30 dB over the wavelength range of 1500–1600 nm.

Schematic diagrams of the structures. (a) The mode/polarization (DE) multiplexer; (b) the ADC for TM modes; (c) the asymmetric-adiabatic-coupler for TE modes; (d) the PSR/PR.
Key parameters of couplers for four higher order modes.
Parameters | w b1 (μm) | w b2 (μm) | w a1 (μm) | w a2 (μm) | w g (μm) | L (μm) |
---|---|---|---|---|---|---|
TM2 | 1.695 | 0.4 | 0.3 | 5.7 | ||
TM1 | 1.092 | 0.43 | 0.3 | 5 | ||
TE2 | 1.04 | 0.93 | 0.26 | 0.32 | 0.18 | 50 |
TE1 | 0.62 | 0.44 | 0.2 | 0.29 | 0.18 | 40 |

Light propagation of TM2(a), TM1(b), TE2(c), and TE1(d) in the corresponding coupler at 1550 nm wavelength.

Calculated transmission of the launched TM2 (a), TM1 (b), TE2 (c), and TE1 (d) modes in the corresponding coupler in the wavelength range of 1500–1600 nm.
2.2 Polarization splitter rotators (PSRs) and polarization rotators (PRs)
The PSR works on the principle of mode hybridness of the ridge waveguide [41, 58], and the slab thickness h s of the ridge waveguide is chosen as 70 nm to be compatible with the standard processes provided by the foundry. Figure 3(c) schematically shows the designed PSR, which contains two parts. One is the polarization rotation region (RI) and the other is the mode splitting region (RII). In region RI, the top ridge is linearly tapered from w b1 to w b2 while the bottom ridge is accordingly tapered from w b1 to w s with a length of L 1. As a result, the TM0 mode can be converted to the TE1 mode when it passes through region R1 due to the mode hybridness, while the TE0 mode passes through directly with an ultra-low loss. In region RII, the TE1 mode in the wide-core couples to the TE0 mode in the narrow core access waveguide by using an asymmetric adiabatic coupler, while the TE0 mode passes through this region without any notable coupling. The key parameters for PSR are summarized in Table 2 [59]. Figure 6(a) and (b) show the simulated light propagation in region RI for the launched TE0 and TM0 modes. One can find that the TE0 mode passes through region RI with negligible losses, while the TM0 mode is converted to the TE1 mode effectively. The corresponding transmission spectra over the wavelength range of 1500–1600 nm are shown in Figure 6(c) and (d). The TE0 mode has an EL less than 0.01 dB, and the TM0-TE1 mode conversion has an EL less than 0.08 dB as well as crosstalk less than −29 dB. Figure 7(a) and (b) shows the simulated light propagation of the TE0 and TE1 modes in region RII. It can be seen that the TE0 mode transmits through the bus waveguide directly and the TE1 mode couples to the TE0 mode in the narrow waveguide, as expected. From the calculated transmission spectra shown in Figure 7(c) and (d), one can find that the TE0 mode has an EL of ∼0.01 dB and the TE1 mode couples to the TE0 mode of the narrow waveguide with an EL less than 0.02 dB as well as crosstalk less than −34 dB. With the combination of regions RI and RII, the TE0 and TM0 modes are finally converted to the TE0 modes at the two output waveguides. The polarization rotators (PRs) to rotate the TM0 mode to the TE0 mode are developed by idling the through port of the PSR.
The key parameters of the PSR.
Parameters | h s | w b1 | w b2 | w b3 | w s | w a1 | w b3 | w g | L 1 | L 2 |
---|---|---|---|---|---|---|---|---|---|---|
Value (μm) | 0.07 | 0.42 | 0.72 | 0.42 | 2.0 | 0.18 | 0.3 | 0.18 | 130 | 60 |

The simulated light propagation of the TE0 (a) and TE1 (b) modes in region RI. The calculated transmissions of the TE0 (c) and TE1 (d) modes in region RI.

The simulated light propagation of the TE0 (a) and TE1 (b) modes in region RII. The calculated transmissions of the TE0 (c) and TE1 (d) modes in region RII.
Summary of the reported silicon ROADM.
Year | Architecture | Channels | Transmission EL (dB) | Add/drop EL (dB) | Crosstalk (dB) |
---|---|---|---|---|---|
(mode × polarization × wavelength) | |||||
2013 [49] | Parent-sub MRR | 4 × 1 × 1 = 4 | 2–4 | ∼2 | −18 ∼ −20 |
2016 [50] | Cascaded MRRs | 1 × 1 × 8 = 8 | 15–18 | 11–15 | −8 ∼ −19 |
2016 [51] | Cascaded MRRs | 1 × 1 × 12 = 12 | 3 | 0–2.5 | −30 |
2016 [52] | AWG + MZS | 1 × 1 × 8 = 8 | ∼15 | ∼7 | −12 ∼ −18 |
2016 [54] | MUX + MZS | 4 × 1 × 1 = 4 | 1–5 | – | −15 ∼ −20 |
2017 [55] | MUX + MRR | 4 × 1 × 1 = 4 | – | 2–5 | ∼−20 |
2020 [56] | MUX + MRR | 3 × 1 × 1 = 3 | ∼7 | ∼10.7 | ∼−13.7 |
2020 [53] | ADC + Bragg | 1 × 1 × 2 = 2 | 1–2.5 | 1–2.5 | −12.8 ∼ −21.4 |
This work | MUX + PSR + MRR | 3 × 2 × 16 = 96 | ∼20 | ∼18 | <−12 |
2.3 MRR-based wavelength-selective switches
Figure 9(a) shows the schematic configuration of MRR-based wavelength-selective optical switch. Two ports of the bus waveguide are used for the straight-forward transmission, and the two cross ports are used for adding and dropping light signals. The cross-section of the MRR waveguide is shown in Figure 9(b), where the metal heater is deposited on the silica upper-cladding along the MRR waveguide to tune the resonant wavelength. The widths of the bus waveguide and the core waveguide are chosen to w b = 0.4 μm, w c = 0.5 μm according to the phase-matching condition, and the waveguide gap is w g = 0.19 μm. The MRRs are designed with a bending radius R varied by following the rule of R = 3.1771 + (n − 1)·0.0058 μm (where n = 1, … 16), in which way their resonant wavelengths increase with a channel spacing of 1.6 nm in the 1550 nm wavelength-band. Definitely a smaller bending radius ensures a larger free spectral range (FSR) for the MRR. Here we choose the bending radius R = 3.1771 μm, so that the FSR is as large as 27 nm to cover more than 16 channels. Figure 8(c) shows the calculated spectra for the drop port of the 16 MRRs in cascade.

The MRR-based wavelength-selective switch. (a) The schematic diagram; (b) the cross-section of the MRR waveguide; (c) The calculated spectra for the drop port of the 16 MRRs in cascade.
3 Fabrication and results
The proposed ROADM was fabricated on an SOI wafer with a 220-nm-thick top-silicon layer and a 2-μm-thick BOX layer. The waveguide was patterned by using an E-beam lithography (EBL) process and formed by a dry-etching process, and then a 1.5-μm-thick SiO2 layer was deposited as the top cladding. Metal microheaters of Cr (20 nm)/Ti (200 nm) alloy were then fabricated by a lift-off process. Finally, a 200 nm thick silica was grown to protect the metal layer, and contact windows were opened upon the contact pad by a dry-etching process. Figure 9(a) shows the microscopy images for the fabricated 96-channels ROADM. A mode MUX and a mode DEMUX are connected to the input and output MBWs of the present ROADM, respectively, so that the mode channels can be characterized selectively. The TM-type and TE-type grating couplers are used for the fiber-chip coupling. For characterization convenience, the six input ports (i. e., I 1–I 6) and six output ports (i. e., O 1–O 6) are placed on the same side, in which way the package with a 12-channel fiber array can be carried out. The enlarged view of the six-channel mode/polarization (DE) MUX and the portion of the MRR-based wavelength-selective switches array are shown in Figure 9(b) and (c). The ROADM chip is bonded to a printed circuit board (PCB) and all 112 pads connected to the metal microheaters on the chip are wire bonded to the PCB for electrical control.

Microscopy images of the fabricated ROADM for hybrid WDM-MDM-PDM systems (a), the enlarged mode/polarization DEMUX (b), and the enlarged view for the MRR-based wavelength-selective optical switches (c).
First the transmission spectra of the fabricated ROADM are measured. Light from a broadband source is launched into the input ports I i (i = 1–6, corresponding to the TM2, TM1, TM0, TE0, TE1, and TE2 mode channels) by a fiber array. Light from the six output ports O j (j = 1–6) is then received with the same fiber array, which connects with an optical spectrum analyzer. Figure 10(a–f) show the measured transmission T ij from port I i to port O j . Here the transmission is normalized with respect to that for a straight waveguide fabricated on the same chip, when i = 1, 2, 3, 4, 5, and 6, respectively. The transmission T ij is mainly dependent on the performance of the pairs of mode/polarization (de)multiplexers and the 6 × 16 MRR array.

Measured transmission T ij at the through port O j (j = 1, 2, 3, 4, 5, 6) when light is inputted into (a) I 1; (b) I 2; (c) I 3; (d) I 4; (e) I 5; (f) I 6, respectively.
When j = i, the transmission T ij is ∼20 dB for TE modes and ∼15 dB for TM modes in the wavelength range of 1560–1590 nm, which includes several parts. First, the EL induced by the 0.7-cm-long loop-back waveguide is about 2 dB and 1.3 dB for the TE0 and TM0 modes, respectively, regarding that the propagation losses are ∼3.1 dB/cm and ∼1.8 dB/cm for the TE0 and TM0 modes, respectively. Second, the loss from the additional mode MUX/DEMUX placed at the input/output ends is about 3–4 dB. Accordingly, the EL for the optical signal transmitted to the through port of the fabricated ROADM has an EL of ∼14 dB and ∼10 dB for the TE and TM modes. Note that the waveguide crossing has an EL of ∼0.08 dB, thus the total EL due to the waveguide crossings involved is about 7.6 dB. Fortunately, the EL can be reduced further to ∼0.01 dB/crossing by optimizing the design of waveguide crossings with the particle swarm optimization (PSO) method [60].
When j ≠ i, the measured transmission T ij gives the crosstalk from port I i to port O j . It is shown that the inter-mode crosstalks (T ij –T ii ) are <−20 dB, −20 dB, −20 dB, −10 dB, −10 dB and −10 dB for the launched TM2, TM1, TM0, TE0, TE1 and TE2 mode, respectively. We also characterized the performances of a pair of mode MUXs connected back to back and a PSR fabricated on the same chip. It is shown that a pair of mode (de)multiplexers connected back to back have inter-mode crosstalk less than −20 dB for all six modes, while the PSR has low crosstalk less than −25 dB in the wavelength range of 1530–1600 nm. Moreover, there are some dips at specific wavelengths observed from the measured transmission of T ij , because the MRR-based switches drop the light signals carried by these wavelengths. The measured FSRs of the MRRs are ∼27 nm, which is consistent with the theoretical prediction.
We also measured the spectra T i−m−n at the drop port D m:n (m = 1, … 6, n = 1, … 16) when light is input into port I i . Figure 11(a–f) show the measurement results in the wavelength range of 1560–1590 nm for the cases with port I i (i = 1, 2, 3, 4, 5, 6), respectively. Take the case with port I 1 as an example, which corresponds to the TM2 mode channel, as shown in Figure 11(a). When m = 1 and i = 1, The transmission T 1−1−n describes the signal propagation from the input port I 1 to the drop port D 1:n (n = 1, … 16). The transmissions at the drop ports D 1:n are shown in Figure 11(a), and their resonance wavelengths are consistent with the dips of the transmission of T 11 shown in Figure 10(a). The crosstalk to the other ports D m:n (m ≠ 1) is also shown in Figure 11(a). The inter-mode crosstalk can be evaluated by (T 1−m−n -T 1−1−n ) for a fixed wavelength channel λ n . It is shown that the inter-mode crosstalks are less than −20 dB, −20 dB, −20 dB, −15 dB, −13 dB, and −12 dB for the TM2, TM1, TM0, TE0, TE1, and TE2 mode-channels, respectively. The fabrication error of the mode (DE)MUX induces the increased crosstalk for the TE2 and TE1 mode channels. It can be seen that the peak power of T i−i−n becomes lower as n increases, because more MRRs and waveguide crossings are involved. As it might be noticed from Figure 11(a)–(f), for mode MUXes used here, the intermode crosstalk is usually low and much more wavelength-dependent than the excess loss. As a result, when the ith mode channel is launched, the envelope of the transmissions T i−m−n (i ≠ m) at the drop ports D m−n (n = 1, … 16) are fluctuant randomly. When injecting optical signals into the add port A m−n , one can add the data to the arbitrary mode- and wavelength-channel in the MBW by using the specific MRR-based wavelength-selective switch.

Measured transmissions T i−m−n at drop port D m−n (m = 1, 2, 3, 4, 5, 6) when light is input from port I i . (a) i = 1; (b) i = 3; (c) i = 3; (d) i = 4; (e) i = 5; (f) i = 6.
Although the channel spacing of the 16 wavelength channels is not uniform due to the random fabrication deviations of MRRs, one can obtain uniform drop spectra by finely tuning the microheater of the MRR-based switch to be aligned with the grids of WDM systems. Figure 12(a) shows the tuning characteristics of an MRR-based wavelength selective switch. The resonant wavelength of the MRR-based switch has a 1.6 nm red-shift when the voltage applied on the heater increases from 0 to 0.8 V (i.e., ∼4.5 mW), which is sufficient to achieve the wavelength-selective adding/dropping. Figure 12(b) shows the switching response of the MRR-based switch, which has a switching rise-time of 17 μs and a drop-time of 8 μs. Here we take the transmission T 1−1−n (n = 1, … 16) as an example to show the tunability of the MRR-based switch. Figure 12(c) and (d) shows the measured drop spectra of the 16 MRRs in cascade without and with thermal tuning, respectively. The channel spacing between these MRRs is tuned to be 1.6 nm by applying appropriate power for heating. When the resonant wavelength of the MRR-based switch is aligned to the target wavelength channel λ n , the signal carried by this wavelength can be dropped from or added to the SMW. However, when the resonant wavelength of MRRs is aligned to λ n + 0.5 Δλ ch, which locates at the central position between two adjacent channels, the optical signal carried by the channel λ n passes through the switches array directly without affecting the other wavelength channels almost. Figure 12(a) shows the 0.8 nm and 1.6 nm inter-wavelength crosstalk of the fabricated MRRs are less than −19 dB and −24 dB. As a result, one can selectively add/drop any mode- and wavelength-channel with low crosstalk by tuning the heating power.

Measured transmissions (a) and switching responses (b) of the MRR-based wavelength-selective switch operating with different heating voltages. The measured drop spectra of the 16 MRR-based wavelength-selective switch without (c) and with (d) heater tuning.
We further characterized the system transmission performances of the present on-chip ROADM, where 10-GBaud quadrature phase shift keying (QPSK) signals are used and the configuration of the experimental setup is shown in Figure 13. For the transmitter, the optical carrier from an external cavity laser (ECL) at a specific wavelength was divided into two paths, one for generating the signal and the other for the local oscillator (LO). The QPSK signals are generated by the I/Q modulator driven by an arbitrary waveform generator (AWG) operated at 10-Gbaud. Then, the optical signals were amplified by an erbium-doped fiber amplifier (EDFA) and coupled into the fabricated ROADM chip by the grating couplers. After passing through the ROADM chip, the optical signals were coupled out of the chip and sent into the receiver, where the received QPSK signals and LO are mixed for coherent detection. Finally, the bit-error-rate (BER) performance was evaluated by digital offline processing after recording the received waveforms with a real-time sampling oscilloscope (OSC) operating at 80 GSa/s.

Experiment setup for the system transmission characterization with an external cavity laser (ECL), a 3-dB coupler, two polarization controllers (PCs), an arbitrary waveform generator (AWG), an I/Q modulator (Mod.), three erbium-doped fiber amplifiers (EDFA), a variable optical attenuator (VOA), a coherent detection receiver, an oscilloscope (OSC), and an offline digital signal processor (DSP).
Figure 14 shows the measured BER as a function of the received optical signal-to-noise ratio (OSNR) for different channels of mth modes and nth wavelengths (C m:n ). Here six channels C6:1, C5:2, C4:4, C3:8, C2:12, and C1:18 are randomly measured and shown in Figure 14(a–f), respectively. Comparing the BER performances of the add/drop transmissions to that of the back-to-back transmission of each channel, the observed OSNR penalties induced by the adding/dropping route of the ROADM are less than 2 dB for all six channels at a BER of 3.8 × 10−3 (forward error correction (FEC) threshold). The insets of Figure 14(a–f) show the measured constellations of QPSK signal for the adding and dropping transmission, respectively, showing decent signal quality. WDM transmission performances are further characterized, where two adjacent wavelengths channels C6:1, C6:2 with a channel spacing of 1.6 nm are used here as an example. As shown in Figure 15(a), the observed OSNR penalties of the dropping transmission at a BER of 3.8 × 10−3 for both channels are less than 1 dB because of the low inter-wavelength crosstalk (<−24 dB) of two adjacent wavelength channels of the fabricated ROADM. The constellations of the QPSK signal for C6:1, C6:2 drop transmission and back-to-back transmission are shown in Figure 15(b). It should be mentioned, that the baud rate of the system is limited by the narrow bandwidth of the MRR switches, which can be improved by adopting higher-order MRRs with box-like responses [61, 62].

Measured BER performances of ROADM for mth modes and nth wavelengths (C m:n ), (a) m = 6, n = 1; (b) m = 5, n = 2; (c) m = 4, n = 4; (d) m = 3, n = 8; (e) m = 2, n = 12; (f) m = 1, n = 16. Insets are constellations of QPSK for the dropping and adding transmissions.

Measured BER performance (a) and the corresponding constellations of QPSK (b) of the WDM dropping transmissions and back-to-back transmission for two adjacent wavelength-channels C6:1 and C6:2.
Table 3 summarizes the reported silicon ROADMs for WDM, MDM, and hybrid-multiplexed systems. It can be seen that the present ROADM is the first one for hybrid MDM-PDM-WDM systems, working with the largest channel number. On the other hand, the excess loss and the crosstalk increase as the channel number increases. As a result, it is really necessary to develop silicon photonic devices with very high performances, e. g., low-loss waveguide crossings, low-crosstalk PSRs, higher-order MRRs with box-like responses, etc. Furthermore, more efforts should be made to include more functional elements, such as online optical power monitors and variable optical attenuators for channel equalization.
4 Conclusions
In summary, we have proposed and demonstrated a silicon-based 96-channel ROADM for hybrid MDM-WDM-PDM systems for the first time. This ROADM consists of a pair of six-channel mode (DE) multiplexers to support six mode channels of dual polarizations, a 6 × 16 array of MRR-based wavelength-selective switches, and 6 × 96 single-mode waveguide crossings for on-chip crossing connection. Such a ROADM integrates more than 1000 elements monolithically, containing 96 MRRs, 576 waveguide crossings, 192 grating couplers, 96 micro-heaters, 112 pads, six PSRs, four asymmetric adiabatic couplers and four asymmetric directional couplers. One can add/drop optical signal to/from any channel carried by any wavelength, any mode and any polarization in the multimode bus waveguide by switching the corresponding MRR-based optical switch. The fabricated ROADM chip was bonded to a PCB for electrical controlling and packaged with a fiber array for light coupling. The ROADM shows on-chip ELs of 5–20 dB, and inter-wavelength crosstalk <−24 dB, and inter-mode crosstalk <−12 dB for adding any one of all 96 channels in the wavelength range of 1560–1590 nm. The system experiments have been demonstrated by using 10-GBaud QPSK signals, showing the observed OSNR penalties induced by the ROADM are less than 2 dB at a BER of 3.8 × 10−3. The performance of the proposed ROADM can be further improved by adopting low loss waveguide crossings, and low crosstalk mode MUX/DEMUX. Such hybrid WDM-MDM-PDM ROADM can be extended with more mode channels as well as more wavelength channels for the applications of ultra-high-capacity data routing/switching, which is promising in future optical interconnect networks.
Funding source: National Natural Science Foundation of China
Award Identifier / Grant number: 61961146003
Award Identifier / Grant number: 62005238
Award Identifier / Grant number: 62125503
Award Identifier / Grant number: 91950205
Funding source: National Major Research and Development Program
Award Identifier / Grant number: 2019YFB2203600
Funding source: National Science Fund for Distinguished Young Scholars
Award Identifier / Grant number: 61725503
Funding source: The Fundamental Research Funds for the Central Universities
Funding source: Zhejiang Provincial Natural Science Foundation
Award Identifier / Grant number: LD19F050001
Funding source: The Leading Innovative and Entrepreneur Team Introduction Program of Zhejiang
Award Identifier / Grant number: 2021R01001
-
Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.
-
Research funding: National Major Research and Development Program (No. 2019YFB2203600); National Science Fund for Distinguished Young Scholars (61725503); National Natural Science Foundation of China (NSFC) (62125503, 91950205, 61961146003, 62005238); Zhejiang Provincial Natural Science Foundation (LD19F050001); The Fundamental Research Funds for the Central Universities; The Leading Innovative and Entrepreneur Team Introduction Program of Zhejiang (2021R01001).
-
Conflict of interest statement: The authors declare no conflicts of interest.
References
[1] T. Moriaka, Y. Awaji, R. Ryf, et al.., “Enhancing optical communications with brand new fibers,” IEEE Commun. Mag., vol. 50, p. S31, 2012. https://doi.org/10.1109/mcom.2012.6146483.Search in Google Scholar
[2] G. G. Li, N. Bai, N. B. Zhao, and C. Xia, “Space-division multiplexing: the next frontier in optical communication,” Adv. Opt. Photonics, vol. 6, p. 413, 2014. https://doi.org/10.1364/aop.6.000413.Search in Google Scholar
[3] C. L. Li, D. J. Liu, and D. Dai, “Multimode silicon photonics,” Nanophotonics, vol. 8, p. 227, 2018. https://doi.org/10.1515/nanoph-2018-0161.Search in Google Scholar
[4] Y. Fazea and V. Mezhuyev, “Selective mode excitation techniques for mode-division multiplexing: a critical review,” Opt. Fiber Technol., vol. 45, p. 280, 2018. https://doi.org/10.1016/j.yofte.2018.08.004.Search in Google Scholar
[5] D. X. Dai and S. P. Wang, “Asymmetric directional couplers based on silicon nanophotonic waveguides and applications,” Front. Optoelectron., vol. 9, p. 450, 2016. https://doi.org/10.1007/s12200-016-0557-8.Search in Google Scholar
[6] D. Gonzalez-Andrade, J. G. Wanguemert-Perez, A. Velasco, et al.., “Ultra-broadband mode converter and multiplexer based on sub-wavelength structures,” IEEE Photon. J., vol. 10, p. 1, 2018. https://doi.org/10.1109/jphot.2018.2819364.Search in Google Scholar
[7] X. Jiang, H. Wu, and D. Dai, “Low-loss and low-crosstalk multimode waveguide bend on silicon,” Opt. Express, vol. 26, p. 17680, 2018. https://doi.org/10.1364/oe.26.017680.Search in Google Scholar
[8] C. L. Sun, Y. Yu, G. Y. Chen, and X. L. Zhang, “Ultra-compact bent multimode silicon waveguide with ultralow inter-mode crosstalk,” Opt. Lett., vol. 42, p. 3004, 2017. https://doi.org/10.1364/ol.42.003004.Search in Google Scholar
[9] C. H. Yao, Z. Wang, H. W. Wang, Y. He, Y. Zhang, and Y. Su, “On-chip multi-mode manipulation via two-dimensional refractive-index perturbation on a waveguide,” Adv. Opt. Mater., vol. 8, p. 2000996, 2020. https://doi.org/10.1002/adom.202000996.Search in Google Scholar
[10] J. Xiang, Z. Tao, X. Li, et al.., “Metamaterial-enabled arbitrary on-chip spatial mode manipulation,” Light Sci Appl., vol. 11, p. 168, 2022. https://doi.org/10.1038/s41377-022-00859-9.Search in Google Scholar PubMed PubMed Central
[11] W. Chang, L. Lu, X. Ren, et al.., “Ultracompact dual-mode waveguide crossing based on subwavelength multimode-interference couplers,” Photonics Res., vol. 6, p. 660, 2018. https://doi.org/10.1364/prj.6.000660.Search in Google Scholar
[12] H. Xu and Y. Shi, “Metamaterial-based Maxwell’s fisheye lens for multimode waveguide crossing,” Laser Photonics Rev., vol. 12, p. 1800094, 2018. https://doi.org/10.1002/lpor.201800094.Search in Google Scholar
[13] W. Zhao, X. Yi, Y. Peng, L. Zhang, H. Chen, and D. Dai, Silicon multimode waveguide crossing based on anisotropic subwavelength gratings. Laser Photonics Rev., 2022, vol. 16, Art. no. 2100623. https://doi.org/10.1002/lpor.202100623.Search in Google Scholar
[14] Q. Huang and K. S. Chiang, “Polarization-Insensitive ultra-broadband mode filter based on 3D graphene structure buried in optical waveguide,” Optica, vol. 7, pp. 744–745, 2020. https://doi.org/10.1364/optica.396918.Search in Google Scholar
[15] H. Xu and Y. Shi, “Ultra-broadband dual-mode 3 dB power splitter based on a Y-junction assisted with mode converters,” Opt. Lett., vol. 41, p. 5047, 2016. https://doi.org/10.1364/ol.41.005047.Search in Google Scholar PubMed
[16] C. C. Ye, M. Zhang, and D. D. Dai, “Broadband dual-mode 2 × 2 3 dB multimode interference couplers with a shallowly etched multimode section,” Applied Opt., vol. 59, pp. 7308–7312, 2020. https://doi.org/10.1364/ao.397676.Search in Google Scholar
[17] Y. Hsu, C. Y. Chuang, X. Wu, et al.., “2.6 tbit/s on-chip optical interconnect supporting mode-division-multiplexing and PAM-4 signal,” IEEE Photon. Technol. Lett., vol. 30, p. 1052, 2018. https://doi.org/10.1109/lpt.2018.2829508.Search in Google Scholar
[18] J. Yao, W. Zhang, H. Ghorbani, et al.., “On-chip 4x10 GBaud/s mode-division multiplexed PAM-4 signal transmission,” IEEE J. Quantum Electron., vol. 26, 2020, Art. no. 8302308.10.1109/JSTQE.2020.2964388Search in Google Scholar
[19] D. X. Dai and J. Bowers, “Silicon-based on-chip multiplexing technologies and devices for peta-bit optical interconnects,” Nanophotonics, vol. 3, p. 283, 2014. https://doi.org/10.1515/nanoph-2013-0021.Search in Google Scholar
[20] Y. He, Y. Zhang, H. Wang, L. Sun, and Y. Su, “Design and experimental demonstration of a silicon multi-dimensional (DE)multiplexer for wavelength-mode- and polarization-division (DE)multiplexing,” Opt. Lett., vol. 45, p. 2846, 2020. https://doi.org/10.1364/ol.390015.Search in Google Scholar PubMed
[21] Y. Zhang, R. H. Zhang, Q. M. Zhu, Y. Yuan, and Y. Su, “Architecture and devices for silicon photonic switching in wavelength, polarization and mode,” J. Light. Technol., vol. 38, p. 215, 2020. https://doi.org/10.1109/jlt.2019.2946171.Search in Google Scholar
[22] J. Wang, S. Chen, and D. Dai, “Silicon hybrid demultiplexer with 64 channels for wavelength/mode-division multiplexed on-chip optical interconnects,” Opt. Lett., vol. 39, p. 6993, 2014. https://doi.org/10.1364/ol.39.006993.Search in Google Scholar
[23] D. X. Dai, J. Wang, S. Chen, S. Wang, and S. He, “Monolithically integrated 64-channel silicon hybrid demultiplexer enabling simultaneous wavelength- and mode-division-multiplexing,” Laser Photonics Rev., vol. 9, p. 339, 2015. https://doi.org/10.1002/lpor.201400446.Search in Google Scholar
[24] S. Wang, H. Wu, M. Zhang, et al.., “A 32-channel hybrid wavelength-/mode-division (DE)multiplexer on silicon,” IEEE Photon. Technol. Lett., vol. 30, p. 1194, 2018. https://doi.org/10.1109/lpt.2018.2839533.Search in Google Scholar
[25] L. W. Luo, N. Ophir, C. P. Chen, et al.., “WDM-compatible mode-division multiplexing on a silicon chip,” Nat. Commun., vol. 5, p. 3069, 2014. https://doi.org/10.1038/ncomms4069.Search in Google Scholar PubMed
[26] B. Stern, X. Zhu, C. P. Chen, et al.., “On-chip mode-division multiplexing switch,” Optica, vol. 2, p. 530, 2015. https://doi.org/10.1364/optica.2.000530.Search in Google Scholar
[27] Y. H. Dong, Y. Zhang, J. Shen, Z. H. Xu, X. H. Zou, and Y. K. Su, “Silicon-integrated high-speed mode and polarization switch-and-selector,” J. Semicond., vol. 43, p. 022301, 2022. https://doi.org/10.1088/1674-4926/43/2/022301.Search in Google Scholar
[28] X. Cao, S. Zheng, N. Zhou, J. Zhang, and J. Wang, “On-chip multi-dimensional 1 × 4 selective switch with simultaneous mode-/polarization-/wavelength-division multiplexing,” IEEE J. Quantum Electron., vol. 56, p. 1, 2020. https://doi.org/10.1109/jqe.2020.3004821.Search in Google Scholar
[29] H. Jia, L. Zhang, J. Ding, L. Zheng, C. Yuan, and L. Yang, “Microring modulator matrix integrated with mode multiplexer and de-multiplexer for on- chip optical interconnect,” Opt. Express, vol. 25, p. 422, 2017. https://doi.org/10.1364/oe.25.000422.Search in Google Scholar PubMed
[30] D. Dai, J. Wang, and Y. Shi, “Silicon mode (DE) multiplexer enabling high capacity photonic networks-on-chip with a single-wavelength-carrier light,” Opt. Lett., vol. 38, p. 1422, 2013. https://doi.org/10.1364/ol.38.001422.Search in Google Scholar PubMed
[31] J. Wang, P. Chen, S. Chen, Y. Shi, and D. Dai, “Improved 8-channel silicon mode demultiplexer with grating polarizers,” Opt. Express, vol. 22, p. 12799, 2014. https://doi.org/10.1364/oe.22.012799.Search in Google Scholar PubMed
[32] J. Wang, S. He, and D. Dai, “On-chip silicon 8-channel hybrid (DE)multiplexer enabling simultaneous mode- and polarization-division-multiplexing,” Laser Photonics Rev., vol. 8, p. L18, 2014. https://doi.org/10.1002/lpor.201300157.Search in Google Scholar
[33] D. Dai, C. Li, S. Wang, et al.., “10-channel mode (DE)multiplexer with dual polarizations,” Laser Photonics Rev., vol. 12, p. 1700109, 2017. https://doi.org/10.1002/lpor.201700109.Search in Google Scholar
[34] L. Liu, Y. Ding, K. Yvind, et al.., “Silicon-on-insulator polarization splitting and rotating device for polarization diversity circuits,” Opt. Express, vol. 19, p. 12646, 2011. https://doi.org/10.1364/oe.19.012646.Search in Google Scholar
[35] Y. Ding, L. Liu, C. Peucheret, and H. Ou, “Fabrication tolerant polarization splitter and rotator based on a tapered directional coupler,” Opt. Express, vol. 20, p. 20021, 2021. https://doi.org/10.1364/oe.20.020021.Search in Google Scholar PubMed
[36] Y. Xu and J. Xiao, “Ultracompact and high efficient silicon-based polarization splitter-rotator using a partially-etched subwavelength grating coupler,” Sci. Rep., vol. 6, p. 27949, 2016. https://doi.org/10.1038/srep27949.Search in Google Scholar PubMed PubMed Central
[37] H. Wu, Y. Tan, and D. Dai, “Ultra-broadband high-performance polarizing beam splitter on silicon,” Opt. Express, vol. 25, p. 6069, 2017. https://doi.org/10.1364/oe.25.006069.Search in Google Scholar PubMed
[38] H. N. Xu, D. Dai, and Y. Shi, “Ultra-broadband and ultra-compact on-chip silicon polarization beam splitter by using hetero-anisotropic metamaterials,” Laser Photonics Rev., vol. 13, p. 1800349, 2019. https://doi.org/10.1002/lpor.201970021.Search in Google Scholar
[39] Y. Xiong, D. X. Xu, J. H. Schmid, P. Cheben, S. Janz, and W. N. Ye, “Fabrication tolerant and broadband polarization splitter and rotator based on a taper-etched directional coupler,” Opt. Express, vol. 22, p. 17458, 2014. https://doi.org/10.1364/oe.22.017458.Search in Google Scholar
[40] H. N. Xu, Y. C. Shi, et al.., “Subwavelength-grating-assisted silicon polarization rotator covering all optical communication bands,” Opt. Express, vol. 27, p. 5588, 2019. https://doi.org/10.1364/oe.27.005588.Search in Google Scholar
[41] D. Dai and J. E. Bowers, “Novel concept for ultracompact polarization splitter-rotator based on silicon nanowires,” Opt. Express, vol. 19, p. 10940, 2011. https://doi.org/10.1364/oe.19.010940.Search in Google Scholar PubMed
[42] P. Dong, W. Qian, H. Liang, et al.., “Low power and compact reconfigurable multiplexing devices based on silicon microring resonators,” Opt. Express, vol. 18, p. 9852, 2010. https://doi.org/10.1364/oe.18.009852.Search in Google Scholar PubMed
[43] L. Zhang, H. Zhao, H. Wang, et al.., “Cascading second-order microring resonators for a box-like filter response,” J. Light. Technol., vol. 35, p. 5347, 2017. https://doi.org/10.1109/jlt.2017.2775658.Search in Google Scholar
[44] S. Cheung, T. Su, K. Okamoto, and S. Yoo, “Ultra-compact silicon photonic 512×512 25 GHz arrayed waveguide grating router,” IEEE J. Sel. Top. Quantum Electron., vol. 20, p. 310, 2014. https://doi.org/10.1109/jstqe.2013.2295879.Search in Google Scholar
[45] S. Pathak, P. Dumon, D. T. Van, and W. Bogaerts, “Comparison of AWGs and Echelle gratings for wavelength division multiplexing on silicon-on-insulator,” IEEE Photon. J., vol. 6, p. 1, 2014. https://doi.org/10.1109/jphot.2014.2361658.Search in Google Scholar
[46] P. Pan, J. An, Y. Wang, et al.., “Compact 4-channel AWGs for CWDM and LAN WDM in data center monolithic applications,” Opt. Laser Technol., vol. 75, p. 177, 2015. https://doi.org/10.1016/j.optlastec.2015.07.005.Search in Google Scholar
[47] S. Gringeri, B. Basch, V. Shukla, et al.., “Flexible architectures for optical transport nodes and networks,” IEEE Commun. Mag., vol. 48, p. 40, 2010. https://doi.org/10.1109/mcom.2010.5496877.Search in Google Scholar
[48] H. Yan, X. Feng, D. Zhan, et al.., “Compact optical add-drop multiplexers with parent-sub ring resonators on SOI substrates,” IEEE Photon. Technol. Lett., vol. 25, p. 1462, 2013. https://doi.org/10.1109/lpt.2013.2267232.Search in Google Scholar
[49] D. Wu, Y. Wu, Y. Wang, et al.., “Reconfigurable optical add-drop multiplexer based on thermally tunable micro-ring resonators,” Opt. Commun., vol. 367, p. 44, 2016. https://doi.org/10.1016/j.optcom.2016.01.018.Search in Google Scholar
[50] V. Sorianello, G. D. Angelis, T. Cassese, et al.., “Polarization insensitive silicon photonic ROADM with selectable communication direction for radio access networks,” Opt. Lett., vol. 41, p. 5688, 2016. https://doi.org/10.1364/ol.41.005688.Search in Google Scholar
[51] S. Chen, Y. Shi, S. He, et al.., “Compact eight-channel thermally reconfigurable optical add/drop multiplexers on silicon,” IEEE Photon. Technol. Lett., vol. 28, p. 1874, 2016. https://doi.org/10.1109/lpt.2016.2574459.Search in Google Scholar
[52] H. Yu, X. Wang, H. Qiu, et al.., “Hitless and gridless reconfigurable optical add-drop (de)multiplexer based on looped waveguide sidewall Bragg gratings on silicon,” Opt. Express, vol. 28, p. 14461, 2020. https://doi.org/10.1364/oe.390218.Search in Google Scholar PubMed
[53] S. P. Wang, W. Hao, H. K. Tsang, et al.., “Monolithically integrated reconfigurable add-drop multiplexer for mode-division-multiplexing system,” Opt. Lett., vol. 41, p. 5298, 2016. https://doi.org/10.1364/ol.41.005298.Search in Google Scholar
[54] S. P. Wang, L. Xiang, S. Gao, et al.., “On-chip reconfigurable optical add-drop multiplexer for hybrid wavelength/mode-division-multiplexing systems,” Opt. Lett., vol. 42, p. 2802, 2017. https://doi.org/10.1364/ol.42.002802.Search in Google Scholar PubMed
[55] Y. Jiang, Z. Zhang, J. Yang, et al.., “A flexible and reconfigurable optical add-drop multiplexer for mode division multiplexing systems,” IEEE Photon. Technol. Lett., vol. 32, p. 99, 2020. https://doi.org/10.1109/lpt.2020.3033714.Search in Google Scholar
[56] P. Dumais, D. J. Goodwill, D. Celo, J. Jiang, and E. Bernier, “Three-mode synthesis of slab Gaussian beam in ultra-low-loss in-plane nanophotonic silicon waveguide crossing,” in 2017 IEEE 14th International Conference on Group IV Photonics (GFP), IEEE, 2017, pp. 97–98.10.1109/GROUP4.2017.8082214Search in Google Scholar
[57] D. Dai and M. Zhang, “Mode hybridization and conversion in silicon-on-insulator nanowires with angled sidewalls,” Opt. Express, vol. 23, p. 32452, 2015. https://doi.org/10.1364/oe.23.032452.Search in Google Scholar
[58] D. Dai and H. Wu, “Realization of a compact polarization splitter-rotator on silicon,” Opt. Lett., vol. 41, p. 2346, 2016. https://doi.org/10.1364/ol.41.002346.Search in Google Scholar
[59] W. K. Zhao, R. R. Liu, Y. Y. Peng, et al.., “High-performance silicon polarization switch based on a Mach–Zehnder interferometer integrated with polarization-dependent mode converters,” Nanophotonics, vol. 11, pp. 2293–2301, 2022. https://doi.org/10.1515/nanoph-2022-0022.Search in Google Scholar
[60] Y. Peng, H. Li, and D. Dai, “Compact silicon photonic waveguide crossings with sub-10-mdB loss,” in Asia Communications and Photonics Conference 2021, 2021, paper T4A.130.10.1364/ACPC.2021.T4A.130Search in Google Scholar
[61] F. Morichetti, M. Milanizadeh, M. Petrini, et al.., “Polarization-transparent silicon photonic add-drop multiplexer with wideband hitless tuneability,” Nat. Commun., vol. 12, p. 4324, 2021. https://doi.org/10.1038/s41467-021-24640-5.Search in Google Scholar PubMed PubMed Central
[62] D. J. Liu, L. Zhang, Y. Tan, and D. Dai, “High-order adiabatic elliptical-microring filter with an ultra-large free-spectral-range,” IEEE J. Lightw. Technol., vol. 39, p. 5910, 2021. https://doi.org/10.1109/jlt.2021.3091724.Search in Google Scholar
© 2022 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.
Articles in the same Issue
- Frontmatter
- Review
- Label-free optical biosensors in the pandemic era
- Research Articles
- High-Q asymmetrically cladded silicon nitride 1D photonic crystals cavities and hybrid external cavity lasers for sensing in air and liquids
- Optical rectification and thermal currents in optical tunneling gap antennas
- Tunable optical nonlinearity of indium tin oxide for optical switching in epsilon-near-zero region
- Strong coupling between quasi-bound states in the continuum and molecular vibrations in the mid-infrared
- Hysteresis and balance of backaction force on dielectric particles photothermally mediated by photonic nanojet
- Experimental demonstration of weak chirality enhancement by hybrid perovskite nanocrystals using photonic spin Hall effect
- Picosecond mode switching and Higgs amplitude mode in superconductor-metal hybrid terahertz metasurface
- Integrated thermal emission microchip based on meta-cavity array
- Electrical tunable topological valley photonic crystals for on-chip optical communications in the telecom band
- Local versus bulk circular dichroism enhancement by achiral all-dielectric nanoresonators
- 96-Channel on-chip reconfigurable optical add-drop multiplexer for multidimensional multiplexing systems
- Four- and five-photon upconversion lasing from rare earth elements under continuous-wave pump and room temperature
- Fluorescence-based thermometry for precise estimation of nanoparticle laser-induced heating in cancerous cells at nanoscale
Articles in the same Issue
- Frontmatter
- Review
- Label-free optical biosensors in the pandemic era
- Research Articles
- High-Q asymmetrically cladded silicon nitride 1D photonic crystals cavities and hybrid external cavity lasers for sensing in air and liquids
- Optical rectification and thermal currents in optical tunneling gap antennas
- Tunable optical nonlinearity of indium tin oxide for optical switching in epsilon-near-zero region
- Strong coupling between quasi-bound states in the continuum and molecular vibrations in the mid-infrared
- Hysteresis and balance of backaction force on dielectric particles photothermally mediated by photonic nanojet
- Experimental demonstration of weak chirality enhancement by hybrid perovskite nanocrystals using photonic spin Hall effect
- Picosecond mode switching and Higgs amplitude mode in superconductor-metal hybrid terahertz metasurface
- Integrated thermal emission microchip based on meta-cavity array
- Electrical tunable topological valley photonic crystals for on-chip optical communications in the telecom band
- Local versus bulk circular dichroism enhancement by achiral all-dielectric nanoresonators
- 96-Channel on-chip reconfigurable optical add-drop multiplexer for multidimensional multiplexing systems
- Four- and five-photon upconversion lasing from rare earth elements under continuous-wave pump and room temperature
- Fluorescence-based thermometry for precise estimation of nanoparticle laser-induced heating in cancerous cells at nanoscale