• Advanced Photonics Nexus
  • Vol. 3, Issue 4, 044001 (2024)
Jorge Parra1, Juan Navarro-Arenas1,2, and Pablo Sanchis1,*
Author Affiliations
  • 1Universitat Politècnica de València, Nanophotonics Technology Center, Valencia, Spain
  • 2Universidad de Valencia, Instituto de Ciencia de Materiales (ICMUV), Paterna, Spain
  • show less
    DOI: 10.1117/1.APN.3.4.044001 Cite this Article Set citation alerts
    Jorge Parra, Juan Navarro-Arenas, Pablo Sanchis, "Silicon thermo-optic phase shifters: a review of configurations and optimization strategies," Adv. Photon. Nexus 3, 044001 (2024) Copy Citation Text show less
    (a) Illustration of a TOPS using a metallic heater on top of the waveguide. (b) Cross section of the TOPS. (c) Simulated temperature distribution of the TOPS. (d) Temporal response of the TOPS upon a square electrical signal applied to the heater with (solid blue line) and without (dotted red line) employing pulse pre-emphasis. The considered TOPS comprises a 500 nm×220 nm Si waveguide with a 2 μm×100 nm Ti heater on top. The gap between the waveguide and the heater is 1 μm. The temperature distribution in the cross section was obtained by solving the conductive heat equation using the COMSOL Multiphysics simulation tool. We considered the thermal constants reported in the literature.20" target="_self" style="display: inline;">20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition with a heat transfer coefficient of 5 W/(m2 K) was set on the surface. The temperature of the remaining boundaries was fixed at 293.15 K (cold).
    Fig. 1. (a) Illustration of a TOPS using a metallic heater on top of the waveguide. (b) Cross section of the TOPS. (c) Simulated temperature distribution of the TOPS. (d) Temporal response of the TOPS upon a square electrical signal applied to the heater with (solid blue line) and without (dotted red line) employing pulse pre-emphasis. The considered TOPS comprises a 500  nm×220  nm Si waveguide with a 2  μm×100  nm Ti heater on top. The gap between the waveguide and the heater is 1  μm. The temperature distribution in the cross section was obtained by solving the conductive heat equation using the COMSOL Multiphysics simulation tool. We considered the thermal constants reported in the literature.20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition with a heat transfer coefficient of 5  W/(m2K) was set on the surface. The temperature of the remaining boundaries was fixed at 293.15 K (cold).
    (a) Illustration of a TOPS using a metallic heater on top of the waveguide with thermal isolation by etching the top cladding and buried oxide. (b) Cross section of the free-standing TOPS. (c) Simulated temperature distribution of the free-standing TOPS. The considered TOPS comprises a 500 nm×220 nm silicon waveguide with a 2 μm×100 nm Ti heater on top. The gap between the waveguide and the heater is 1 μm. The temperature distribution in the cross section was obtained by solving the conductive heat equation using the COMSOL Multiphysics simulation tool. We considered the thermal constants reported in the literature.20" target="_self" style="display: inline;">20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition with a heat transfer coefficient of 5 W/(m2 K) was set on the boundaries in contact with air. The temperature of the remaining boundaries was fixed at 293.15 K (cold).
    Fig. 2. (a) Illustration of a TOPS using a metallic heater on top of the waveguide with thermal isolation by etching the top cladding and buried oxide. (b) Cross section of the free-standing TOPS. (c) Simulated temperature distribution of the free-standing TOPS. The considered TOPS comprises a 500  nm×220  nm silicon waveguide with a 2  μm×100  nm Ti heater on top. The gap between the waveguide and the heater is 1  μm. The temperature distribution in the cross section was obtained by solving the conductive heat equation using the COMSOL Multiphysics simulation tool. We considered the thermal constants reported in the literature.20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition with a heat transfer coefficient of 5  W/(m2K) was set on the boundaries in contact with air. The temperature of the remaining boundaries was fixed at 293.15 K (cold).
    (a) Illustration of a TOPS using a transparent heater directly on top of the waveguide. (b) Cross section of the TOPS. (c) Simulated temperature distribution of the TOPS using an ITO heater. The considered TOPS comprises a 500 nm×220 nm silicon waveguide with a 2 μm×100 nm ITO heater on top. The gap between the waveguide and the heater is 100 nm. The temperature distribution in the cross section was obtained by solving the conductive heat equation using the COMSOL Multiphysics simulation tool. We considered the thermal constants reported in the literature.20" target="_self" style="display: inline;">20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition with a heat transfer coefficient of 5 W/(m2 K) was set on the surface. The temperature of the remaining boundaries was fixed at 293.15 K (cold).
    Fig. 3. (a) Illustration of a TOPS using a transparent heater directly on top of the waveguide. (b) Cross section of the TOPS. (c) Simulated temperature distribution of the TOPS using an ITO heater. The considered TOPS comprises a 500  nm×220  nm silicon waveguide with a 2  μm×100  nm ITO heater on top. The gap between the waveguide and the heater is 100 nm. The temperature distribution in the cross section was obtained by solving the conductive heat equation using the COMSOL Multiphysics simulation tool. We considered the thermal constants reported in the literature.20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition with a heat transfer coefficient of 5  W/(m2K) was set on the surface. The temperature of the remaining boundaries was fixed at 293.15 K (cold).
    (a) Illustration of a TOPS utilizing a silicon-doped heater, where the heat generation occurs within the doped silicon waveguide. In this configuration, the waveguide is of the rib type, with several silicon-doped heaters arranged in electrical parallel to minimize total resistance. Metallic contacts are linked to the silicon waveguide via silicon-doped strips. (b) Simulated temperature distribution within the TOPS, consisting of a 500 nm×220 nm silicon waveguide atop a 100-nm-thick slab, with 1 μm-thick SiO2 cladding. Temperature distribution analysis was performed by solving the conductive heat equation with the COMSOL Multiphysics simulation tool, considering the waveguide core as the heat source, based on thermal constants from the literature.20" target="_self" style="display: inline;">20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition, with a heat transfer coefficient of 5 W/(m2 K), was applied on the surface, while the temperature for all other boundaries was fixed at 293.15 K (cold). (c), (d) Cross-sectional views of the TOPS featuring (c) direct current injection and (d) a pn junction setup.
    Fig. 4. (a) Illustration of a TOPS utilizing a silicon-doped heater, where the heat generation occurs within the doped silicon waveguide. In this configuration, the waveguide is of the rib type, with several silicon-doped heaters arranged in electrical parallel to minimize total resistance. Metallic contacts are linked to the silicon waveguide via silicon-doped strips. (b) Simulated temperature distribution within the TOPS, consisting of a 500  nm×220  nm silicon waveguide atop a 100-nm-thick slab, with 1  μm-thick SiO2 cladding. Temperature distribution analysis was performed by solving the conductive heat equation with the COMSOL Multiphysics simulation tool, considering the waveguide core as the heat source, based on thermal constants from the literature.20 A nonuniform tetrahedral mesh, with element sizes ranging from 1 to 500 nm, was employed. A conductive heat flux boundary condition, with a heat transfer coefficient of 5  W/(m2K), was applied on the surface, while the temperature for all other boundaries was fixed at 293.15 K (cold). (c), (d) Cross-sectional views of the TOPS featuring (c) direct current injection and (d) a pn junction setup.
    (a) Illustration of a TOPS using folded waveguides based on a spiral waveguide with a wide heater on top. (b) Cross section of the folded TOPS. The folded waveguide needs to be designed to avoid cross-coupling between adjacent waveguides.
    Fig. 5. (a) Illustration of a TOPS using folded waveguides based on a spiral waveguide with a wide heater on top. (b) Cross section of the folded TOPS. The folded waveguide needs to be designed to avoid cross-coupling between adjacent waveguides.
    (a) Illustration of a TOPS utilizing a multimode waveguide where light is recycled N times through a multipass structure, demonstrating how power consumption decreases as the number of passes increases. (b) Cross section of the TOPS within the multimode waveguide. (c) Depiction of optical mode conversion as a function of the multipass structure’s length. Light enters the structure in the fundamental mode and, after N passes, is converted to the Nth-order mode before being output from the structure and reverted to the fundamental mode.
    Fig. 6. (a) Illustration of a TOPS utilizing a multimode waveguide where light is recycled N times through a multipass structure, demonstrating how power consumption decreases as the number of passes increases. (b) Cross section of the TOPS within the multimode waveguide. (c) Depiction of optical mode conversion as a function of the multipass structure’s length. Light enters the structure in the fundamental mode and, after N passes, is converted to the Nth-order mode before being output from the structure and reverted to the fundamental mode.
    Ref.Structure/heater metalOptimization strategyLoss (dB)Pπ(mW)Switching time (μs)aFOM (mW μs)Length (μm)
    21MZI/CrAuNone32b503.5175700
    22MZI/N/ANone12b2356014×1032500
    23MZI/N/ANone22b901009000140
    24MRR/TiAir trenchesN/A101010030
    25MZI/PtNone16b4030120040
    26MRR/NiPulse pre-emphasis<1164c /<1d64c /<16d60
    27MZI/PtFree-standing2.8b0.5414176100
    28MRR/TiFree-standing<11.217020450
    29MZI/NiSiClose heater<120360200
    30MZI/TiNFree-standing<10.49144711000
    31MRR/NiSiFree-standing<12.9358103850
    32Microdisk NiCrClose heater and pulse pre-emphasis<1122.9c/0.085d35c/1d60
    33MZI/WNone<12245990200
    19MZI/TiNGeometry<1307210320
    Table 1. Summary of basic experimental TOPSs using metallic heaters in SiPh.
    Ref.Structure/heater materialGap (nm)Loss (dB)Pπ (mW)Switching time (μm)aFOM (mWμs)Length (μm)
    35MZI/metal + graphene05b>5020>1000120
    35Microdisk/graphene0<123.513305.55
    36MRR/graphene240<1113.538.555
    37MZI-PhCW/graphene111.12<1<220
    38PhCC/graphene02N/A1.5N/A5
    39MRR/CNTs0N/A14.54.565.3315
    40MRR/MoS2300.427.525187.5283
    20MZI/ITO6600.019.75.250.4450
    41MZI/IHO00.59.60.989.4110
    Table 2. Summary of basic experimental TOPSs using transparent heaters in SiPh.
    Ref.StructureDopant/concentrationCurrent injectionLoss (dB)Pπ(mW)Switching time (μm)aFOM (mW  μs)Length (μm)
    44MZIp-type (B)/1018  cm3Direct360.63.6115
    45MRRn-type (As)/1.8×1018  cm3Direct0.512.72.430.510
    46MZIp-type (B)/7×1017  cm3Direct0.22537561.6
    47MZIn-type (N/A)/N/AParallel heatersN/A255125100
    48MRRp- and n-type (N/A)/2×1018 and 4×1017  cm3Direct with pn junctionN/A19.50.457.8125
    49MRRn-type (As)/4×1013  cm2Direct with pn junction2.5144563.4
    50MZIp- and n-type (N/A)/N/AParallel heater with pn junction1.6b20.997.5200050
    19MZIn-type (P)/1020  cm3Parallel heaters<0.422.82.250.2320
    51MZIp-type (P)/1018  cm3Direct22925835
    52MZIp-type (N/A)/1018  cm3Direct0.2422.60.511.315
    Table 3. Summary of basic experimental TOPSs using doped silicon heaters in SiPh.
    Ref.StructureNumber of foldsLoss (dB)Pπ (mW)Switching time (μm)aFOM (mWμs)Length (μm)
    54MZI596b6.5149113000
    55MZI92.9b4.2c/0.095d65c/1200d237c/114d2900
    56MZI141.232.563589.62300
    57MZI220.9311331876
    Table 4. Summary of advanced experimental TOPSs using folded waveguides and metallic heaters in SiPh.
    TechnologyInsertion lossStatic power consumptionSwitching timeFootprintManufacturability
    Silicon TOPSUltralow (<1 dB)Very high (>mW)Very slow (>μs)Large (>100  μm)Excellent
    Silicon PDEHigh (>1 dB)Moderate (>μW)Very fast (<ns)Very large (mm)Excellent
    MEMSLow (1  dB)Ultralow (nW)Slow (μs)Compact (100  μm)Good
    PlasmonicsVery high (>5 dB)Ultralow (nW)Ultrafast (ps)Ultracompact (μm)Limited
    FerroelectricsUltralow (<1 dB)Ultralow (nW)Ultrafast (ps)Very large (mm)Limited
    PCMsLow (1  dB)ZeroSlow (μs)Ultracompact (μm)Limited
    Table 5. Comparison of mainstream and emerging electro-optic technologies for implementing phase shifters in SiPh.
    Jorge Parra, Juan Navarro-Arenas, Pablo Sanchis, "Silicon thermo-optic phase shifters: a review of configurations and optimization strategies," Adv. Photon. Nexus 3, 044001 (2024)
    Download Citation