Open Access
Add to BibSonomy
Abstract
In this work, programmable optical switching integrated chips for 4-bit binary true/inverse/complement optical code conversions (OCCs) are proposed based on fluorinated photopolymers. Fluorinated bis-phenol-A novolac resin (FAR) with low absorption loss and fluorinated polyacrylate (FPA) with high thermal stability are self-synthesized as core and cladding layer, respectively. The basic architecture of operating unit for the photonic chip designed is composed of directional coupler Mach-Zehnder interferometer (DC-MZI) thermo-optic (TO) switching, X-junction, and Y-bunching waveguide structures. The waveguide module by cascading 16 operating units could realize OCCs function through optical transmission matrix. The response time of the 4-bit binary OCCs is measured as about 300 µs. The insertion loss and extinction ratio of the actual chip are obtained as about 10.5 dB and 15.2 dB, respectively. The electric driving power consumption for OCCs is less than 6 mW. The true/inverse/complement OCCs are achieved by the programmable modulation circuit. The proposed technique is suitable for achieving optical digital computing system with high-speed signal processing and low power consumption.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
With fast development of high-speed cloud computing and large-capacity data center [1–3], photonic integrated circuits (PICs) are attracting great interest and are desired to play key role for high-density optical encryption and processing module [4]. Similar to the classification of electronic computing, optical computing technique could be mainly divided into two categories as optical analog and digital computing [5,6]. Specially, the optical digital computing has been considered as one most feasible way to achieve the technological aim for artificial intelligence photonic processor [7,8]. While, optical code conversions (OCCs) are the fundamental elements for building digital optical computing system [9]. The OCCs applications including storing, computing, and encryption of optical information can be implemented by using various optical switching arrays [10–12]. Individual optical switching unit could be integrated with others to generate logical photonic light-wave circuits for realizing more powerful computing ability [13–15]. Different types of optical waveguide structures including of micro-ring resonators (MRRs), Mach-Zehnder interferometers (MZIs), directional couplers (DCs), and multimode interference devices (MMIs) have been utilized as the basic unit for the optical switching integrated network [16–19]. Several photonic material platforms have been used for achieving programmable optical integrated waveguide chip such as silicon-on-insulator (SOI), lithium niobite (LiNbO3), Indium phosphide (InP) and polymers [20–23]. In particularly, compared to other waveguide material systems, with the greater flexibility, compatibility and processability, polymer optical switching integrated chips are suitable for realizing large-scale, high-density, and low-cost optical calculating processor module [24,25]. The technique is desirable to realize programmable PICs for smart optical communication network and photonic supercomputing applications in future [26–28].
In this work, programmable optical switching integrated chips for 4-bit binary true/inverse/complement OCCs are proposed using fluorinated core and cladding photopolymers. Fluorination feature, optical absorption characteristic, and thermal stability of the waveguide photopolymers are given. The epoxy cross-linking networks generated in the core and cladding photopolymers by UV alignment technology could guarantee the waveguide structure without solvent erosion. The programmable optical switching integrated chips are designed depending on tunable optical encoding transmission matrix approach, which are fabricated by UV direct defining technology based on the fluorinated core and cladding photopolymers. Various OCCs functions of the optical waveguide chips are analyzed and simulated. The programmable operating performances of the chips based on thermo-optic (TO) modulation are measured. The technique is suitable for achieving optical digital computing system with high-speed signal processing and low power consumption.
2. Experiment
2.1 Synthesis and characteristics of Fluorinated core and cladding photopolymers
Materials and methods should be described in sufficient detail to allow the experimental work to be reproduced in another laboratory, and to leave the reader in no doubt as to how the results were derived. To realize the 4-bit binary OCCs waveguide chips, the photopolymers as fluorinated bis-phenol-A novolac resin (FAR) and fluorinated polyacrylate (FPA) are self-synthesized. To gain the FAR, 4,4’ -(Hexafouoro-isopropylidene) diphenol and formaldehyde react under the catalysis of H2SO4 solution, the reaction product and epoxy chloropropane under the catalysis of NaOH solution could generate the FAR. To obtain the FPA, 1 H,1 H,2 H,2H-perfluorooctyl acrylate (PFAC6), trifluoroethyl methacrylate (TFMA), 2-hydroxyethyl methacrylate (HEMA), glycidyl methacrylate (GMA) and styrene are copolymerized to create the FPA using azobisisobutyronitrile (AIBN) as free radical initiator. After that, iodonium salt as photo-initiator is added into both FAR and FPA to initiate epoxy crosslinking polymerization. The FAR and FPA photopolymers are used as core and cladding waveguide materials, respectively. The structural schematic diagrams of the UV cured FAR and FPA photopolymers with epoxy crosslinking networks are shown in Figs. 1(a) and 1(b), respectively. The specific synthesizing procedures for FAR and FPA photopolymers are provided in Supplement 1. In the fluorinated photopolymers, C-F bonds replace of C-H bonds, which could reduce effectively the vibration absorption from the C–H bonds in near-infrared wavelength region. To analyze the fluorinated effectiveness of the photopolymers, the 19F NMR spectra for FAR and FPA are measured and shown in Figs. 2(a) and 2(b), respectively. In Fig. 2(a), it is given that FAR contains F atom, and its characteristic peak is 64.24 ppm. In Fig. 2(b), it is observed that the FPA contains monomer PFAC6 and TFEMA. The characteristic absorption peak of F atoms in TFEMA is 80.01 ppm, and the rest are the characteristic absorption peaks of F atoms in PFAC6. It could be found from the integral area of the F atom that the reaction ratio of the two raw materials is 3:4, which is basically consistent with the feed ratio. As shown in Fig. 2(c), FAR has a lower absorption loss in the 1550 nm wavelength region than the commercial SU-8 material, which is more suitable as a waveguide core layer material. Compared to normal PMMA material, the glass transition temperature (Tg) of FPA increased by 21°C in Fig. 2(d), which suggests that it will have better thermal stability as cladding layer material. To specifically analyze the optical loss of the polymer materials used, the absorption loss of FAR, commercial SU-8, FPA and commercial PMMA materials for different wavelength bands are given as Table 1. It could be observed that as core layer material, the value of FAR is at least one order of magnitude lower than that of commercial SU-8; as cladding layer material, the value of FPA is also relatively better than commercial PMMA. The reason is mainly from the lower absorption loss from the C–F bonds compared with that of C–H bonds in near-infrared wavelength region. These merits of FAR and FPA materials are beneficial to reduce the propagation loss and enhance the operating stability for overall polymer waveguide chips.
Fig. 2. Characterization of materials: The 19F NMR spectra for (a) FAR and (b) FPA; (c) Absorption spectra of FAR and SU-8; (d) Scanning calorimetry curves of FPA and normal PMMA.
Table 1. Absorption losses of FAR, commercial SU-8, FPA, and commercial PMMA materials
2.2 Device structure design and simulation
To realize true/inverse/complement OCCs functions, an operating unit composed of DC-MZI TO switch, X-junction, and Y-bunching is designed, as shown in Fig. 3(a). At 1550 nm wavelength, the refractive indices of FAR and FPA were measured as 1.517 and 1.45 by ellipsometer (SPEL M-2000VI, America), respectively. A 5 µm-thickness SiO2 buffer layer grown on silicon substrate is used as bottom cladding. The core layer of the waveguide is confined to 5 × 5 µm2 and is covered by a top cladding of 8 µm thickness, which can effectively ensure the single-mode optical signal transmission at 1550 nm wavelength, as shown in Fig. 3(b). Then aluminum electrode heaters with 100 nm thickness are set on the active waveguide region. It can be observed that the thermal field generated by the heaters is focused effectively on the active waveguide region in Fig. 3(c).
Fig. 3. (a) The structure diagram of the operating unit; (b) Cross-section optical field distribution of the waveguide; (c) Thermal field distribution for the cross-section with waveguide and electrode heaters; The optimized design of (d) DC-MZI TO switching, (e) X-junction, and (f) Y-bunching.
For the DC-MZI TO switching in the operating unit, the optimized lengths of 3-dB coupler, modulation arms, and directional coupler are defined as 1880, 2000, and 1200 µm, respectively. As shown in Fig. 3(d), there is no phase difference between the two modulation arms at room temperature (Eij ∼ U0), and the signals with the same power are output from O1 and O2 as Splitting state. When the temperature increases by 1.1 °C (Eij ∼ U1), a phase difference of 1/2Lπ is generated between two modulation arms and the signal is focused on O1 as Bar state. As the temperature rises to 3.3 °C (Eij ∼ U2), the signal is focused on O2 due to the phase difference of 3/2Lπ between the two modulation arms as Crossing state. The optical field propagation characteristics for the operating unit are shown in the illustration. For each operating unit, the presence or absence of the input signal can be encoded with Cij as 0 or 1, and the electrode heaters are operated at U0, U1, or U2 by taking 0, 1, or 2 for Eij, respectively. The truth table of operating unit is described in the Table 2.
Table 2. Truth table of operating unit
The X-junction is composed of two bent waveguides with a length of 680 µm and a width of 100 µm. The optical field transmission of the X-junction is shown in the Fig. 3(e). The input signal from A1 will be transferred to B2, while the input signal from A2 will be transferred to B1. The optical field propagation characteristics for the Y-bunching structure with the size of 27 × 3161.5 µm2 is shown in Fig. 3(f). If there are input signals at any port of M1 or M2, it can both be output at N.
The overall schematic structure of the optical integrated waveguide chip for 4-bit binary true/inverse/complement OCCs with 16 cascaded operating units is described in Fig. 4(a). The equivalent optical path of the OCCs waveguide chip is shown in Fig. 4(b). Input and output ports are defined as C1 ∼ C4 and P1 ∼ P4, respectively. The electrode heater modulation state of each operating units is set as E11 ∼ E44.
Fig. 4. The architecture and equivalent optical path of the OCCs waveguide chip: (a) The schematic of the OCCs waveguide chip with waveguide and electrode heater parts; (b) The equivalent optical path of the OCCs waveguide chip.
To achieve OCCs of true, inverse, and complement code between positive and negative numbers, C1 is taken as the binary symbol bit (0 is positive numbers and 1 is negative numbers). According to the input states of C1 ∼ C4, the OCCs of P1 ∼ P4 are shown in the truth Table 3. According to the output states of P1 ∼ P4 at true, inverse, and complement code, the modulation schemes of E11 ∼ E44 are given in Table 4.
Table 3. Truth table of the output states of P1 ∼ P4
Table 4. Truth table of the output states of P1 ∼ P4a
As described in Table 4, it can be found that there is a common electrode heater modulation key matrix of K(Tr)/K(In+)/K(Co+) in the positive true/inverse/complement code and the negative true code, as presented as Eq. (1).
As for the inverse codes of negative, there are two matrices of K1(In-) and K2(In+) determined together, as presented as Eqs. (2) and 3, respectively. As well as the inverse codes of negative, the two matrices with K1(Co-) and K2(Co+) of the negative complement codes are presented as Eqs. (4), (5), and (6), respectively.
To verify OCCs function of programmable optical switching integrated chips, the transmission optical fields are shown in Fig. 5. For positive such as C∼ [0 0 0 1], [0 0 1 1], and [0 1 1 1], the output codes of P are consistent with the true/inverse/complement codes (P = PTr = PIn = PCo), depending on the modulation schemes in Table 3. As for negative such as C∼ [1 1 0 0], [1 1 1 0], and [1 1 1 1], the output codes of P are the same as the true (PTr), inverse (PIn), and complement (PCo) codes, respectively.
Fig. 5. The transmission optical fields of the OCCs waveguide chip simulated by Rsoft software under different modulation schemes.
2.3 Fabrication and measurement of OCCs waveguide chip
The OCCs waveguide chip was fabricated by directly UV-written technique with the detailed process as shown in Fig. 6. Firstly, the FAR core layer with a thickness of 5 µm was spin-coated at 2000 r/s on the surface of SiO2 buffer layer on Si substrate, and prebaked for 10 min at 60 °C and 20 min at 90 °C in Fig. 6(a). The waveguide structures were defined by UV alignment (ABM Co. Inc., USA) under 20 mW/cm2 for 10 s in Fig. 6(b). Next, the chip baked at 120 °C for 20 min was immersed in a special developer (PGMEA) for 15 s to remove the uncrossing-linked core material. Subsequently, the FPA cladding layer with a thickness of 8 µm was spin-coated at 2000 r/s, then prebaked for 20 min at 90 °C, as described in Fig. 6(c). The FPA cladding layer was done at the same UV written process for 120 s and cured at 120 °C for 30 min. Finally, Al electrode heaters with a thickness of 100 nm were deposited onto the FPA cladding by thermal evaporation and done by UV lithography process, as presented in Fig. 6(d).
As shown in Fig. 7(a), the OCCs waveguide chip fixed in the acrylic fixture was connected to the external PCB via a metal spring probe for the contact electrode heater. Both sides of the chip were connected to the 4-channal fiber arrays by a refractive index matching fluid by UV cured. To realize the function of the modulation scheme as shown in Table 3, the programmable modulation circuit designed was shown in Fig. 7(b), which can realize the independent control function of 16 electrode heaters. The top-view of the OCCs waveguide chip with Al electrode heaters is shown in Fig. 7(c). The widths of electrode heaters in contact with the metal spring probe and the heaters on the active region are 2000 and 20 µm, respectively, and the width of linear connection wire between them was 100 µm. The cross-sectional profile of the waveguide was obtained by an optical microscope (×1000), as shown in Fig. 7(d). It could be seen that the size of the waveguide was 5 × 5 µm2, which was consistent with the design.
Fig. 7. Morphologies of the OCCs waveguide chip: (a) The fabricated chip under test package; (b) The programmable modulation circuit; (c) The surface morphology of the fabricated chip with electrode heaters; (d) Optical micrograph of the cross-sectional profile (×1000).
To measure the OCCs waveguide chip, the test system was set up. An infrared laser source (Santec TSL-210) with a central wavelength of 1550 nm was used as the input optical source. Optical signals were selectively sent to each port of the 4-channel fiber arrays (FA) by fiber or beam splitter (1 × 2, 1 × 3, 1 × 4). The control electrical signal from the digital signal generator (SP1642B) is sent to the PCB control board by metal wires, which controlled the operation of electrode heaters of the OCCs waveguide chip. The optical signal from one channel of FA was coupled to the photodetector, converted into the electrical signal. Then, the electrical signal would be compared with the original signal in the digital oscilloscope (DS4024).
A square wave control signal with a frequency of 200 Hz and an amplitude of 5 V was applied to the electrode heater of E41, and the output response curve of the output optical signal at P1 was measured by the digital oscilloscope, as shown in Fig. 8(a). The rising and falling times were 342.9 and 241.4 µs, respectively. The relation curves between output power (P1 and P2) and the driving electrical power consumption (E41) in different operating states was shown in Fig. 8(b). Without the power consumption, the insertion losses of P1 and P2 at Splitting state were 17.5 dB and 17.6 dB, respectively. When the power consumption of the driving electrode heater was 2 mW (E41 ∼ 0) without output signal on the rest of the ports (P3 ∼ 0, P4 ∼0), the insertion loss of P1 and the extinction ratio at Bar state were 10.7 dB and 15.2 dB, respectively. As the power consumption of the driving electrode heater rose to 6 mW (E41 ∼ 1), the insertion loss of P2 and the extinction ratio at Crossing state were 10.5 dB and 15.5 dB, respectively. The near-infrared output optical fields were collected by a charge-coupled device (CCD) camera, corresponding to the modulation schemes in Fig. 5, as shown in Fig. 8(c). It could be seen that the far-field distributions under 12 modulation states were identical with the designs.
Fig. 8. (a) Switch on-off time curves at P1; (b) Extinction ratio with power consumption; (c) The near-infrared output optical fields under different modulation schemes.
As given in Table 5, compared with main characteristics of the optical switching devices based on SOI platform reported before [29–31], it could be found that the insertion losses of these chips are similar, but the proposed programmable optical switching integrated chips based on fluorinated photopolymers in this work has lower electric consumption power and fabrication cost. In addition, depending on the low temperature directly UV-written technique, fluorinated photopolymer platform has faster, more controllable, and simpler process flow to a certain extent. Meanwhile, it could be expected that by interlayer coupling approach and organic-inorganic hybrid integration technology, on-chip photoelectric detecting elements might be loaded on the polymer photonic integrated chip.
Table 5. Compared with switches based on SOI platform in other reports
3. Conclusion
In summary, programmable optical switching integrated chips based on fluorinated photopolymers are proposed for 4-bit binary true/inverse/complement OCCs. FAR and FPA self-synthesized with epoxy-crosslinking networks were applied as optical waveguide core and cladding materials, which effectively reduced absorption loss and improved thermal stability of the chip. The thermal and optical characteristics of the chip were analyzed based on the finite element method, and the architecture based on 16 operating units cascaded was optimized. The rising and falling response times of the 4-bit binary OCCs were obtained as 342.9 and 241.4 µs, respectively. The insertion loss, extinction ratio, and power consumption of OCCs were measured as 10.5 dB, 15.5 dB, and 6 mW, respectively. 4-bit binary true/inverse/complement OCCs of the photonic chip were achieved. The proposed technique is suitable for achieving optical digital computing system with high-speed signal processing and low power consumption.
Funding
National Natural Science Foundation of China (11873058); Jilin Scientific and Technological Development Program (20230101125JC).
Disclosures
The authors declare no conflict of interests related to this article.
Data availability
The data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. S.S. Manvi and G.K. Shyam, “Resource management for Infrastructure as a Service (IaaS) in cloud computing: A survey,” J. Netw. Comput. Appl 41(1), 424–440 (2014). [CrossRef]
2. X.H. You, C.X. Wang, J. Huang, et al., “Towards 6 G wireless communication networks: vision, enabling technologies, and new paradigm shifts,” Sci. China. Inform. Sci 64(1), 110301 (2021). [CrossRef]
3. L.H. Shen, K.T. Feng, and L. Hanzo, “Five Facets of 6G: Research Challenges and Opportunities,” ACM Comput. Surv. Appl 55(11), 1–39 (2023). [CrossRef]
4. Y.X. Hou, Y. Li, Z.C. Zhang, et al., “Large-Scale and Flexible Optical Synapses for Neuromorphic Computing and Integrated Visible Information Sensing Memory Processing,” ACS Nano 15(1), 1497–1508 (2021). [CrossRef]
5. R. Soref, B.J. Shastri, and A.N. Tait, “The Silicon-Based XOI Wafer: The Most General Electronics-Photonics Platform for Computing, Sensing, and Communications,” IEEE J. Sel. Top. Quant 29(2: Optical Computing), 1–8 (2023). [CrossRef]
6. G. George, T. Apostolos, M.P. Miltiadis, et al., “Analog nanophotonic computing going practical: silicon photonic deep learning engines for tiled optical matrix multiplication with dynamic precision,” Nanophotonics 12(5), 963–973 (2023). [CrossRef]
7. B.J. Shastri, A.N. Tait, T.F.D. Lima, et al., “Photonics for artificial intelligence and neuromorphic computing,” Nat. Photonics 15(2), 102–114 (2021). [CrossRef]
8. J. Feldmann, N. Youngblood, M. Karpov, et al., “Parallel convolutional processing using an integrated photonic tensor core,” Nature 589(7840), 52–58 (2021). [CrossRef]
9. K. Inoshita, H. Kishikawa, Y. Makimoto, et al., “Proposal of Optical Waveguide Circuits for Recognition of Optical QAM Codes,” J. Lightwave Technol. 31(13), 2271–2278 (2013). [CrossRef]
10. J. Feldmann, N. Youngblood, X. Li, et al., “Integrated 256 Cell Photonic Phase-Change Memory With 512-Bit Capacity,” IEEE J. Sel. Top. Quant 26(2), 1–7 (2020). [CrossRef]
11. C. Qian, X. Lin, X.B. Lin, et al., “Performing optical logic operations by a diffractive neural network,” Light: Sci. Appl. 9(1), 59 (2020). [CrossRef]
12. F.K. Bian, L.Y. Sun, H.X. Chen, et al., “Bioinspired Perovskite Nanocrystals-Integrated Photonic Crystal Microsphere Arrays for Information Security,” Adv. Sci 9(9), 2105278 (2022). [CrossRef]
13. W. Bogaerts, D. Pérez, J. Capmany, et al., “Programmable photonic circuits,” Nature 586(7828), 207–216 (2020). [CrossRef]
14. Y.W. Xie, S.H. Hong, H. Yan, et al., “Low-loss chip-scale programmable silicon photonic processor,” Opto-Electron. Adv. 6(3), 220030 (2023). [CrossRef]
15. N.C. Harris, J. Carolan, D. Bunandar, et al., “Linear programmable nanophotonic processors,” Optica 5(12), 1623–1631 (2018). [CrossRef]
16. S.L. Pan, Z.Z. Tang, M.H. Huang, et al., “Reflective-Type Microring Resonator for On-Chip Reconfigurable Microwave Photonic Systems,” IEEE J. Sel. Top. Quant 26(5), 1–12 (2020). [CrossRef]
17. S. Singh and Lovkesh, “Ultrahigh Speed Optical Signal Processing Logic Based on an SOA-MZI,” IEEE J. Sel. Top. Quant 18(2), 970–977 (2012). [CrossRef]
18. R. Nanda, R. Rath, S. Swarnakar, et al., “Design of All Optical Directional Coupler Using Plasmonic MIM Waveguide for Switching Applications,” Plasmonics 17(5), 2153–2159 (2022). [CrossRef]
19. M. Tajaldini and M. Z. Jafri, “Simulation of all-optical NOT/OR/AND/NAND/ NOR/XNOR/XOR logic gates using a single structure high contrast index nonlinear multimode interference coupler,” Phys. Scripta 94(10), 105004 (2019). [CrossRef]
20. X.Y. Gao, W.T. Gu, W.C. Dong, et al., “Seven-channel all-optical reconfigurable canonical logic units multicasting at 40 Gb/s based on a nonlinearity-enhanced silicon waveguide,” Opt. Express 30(18), 32650–32659 (2022). [CrossRef]
21. Y.C. Jia, L. Wang, and F. Chen, “Ion-cut lithium niobate on insulator technology: Recent advances and perspectives,” Appl. Phys. Rev 8(1), 1–28 (2021). [CrossRef]
22. L.N. Zhang, Y.R. Xu, J. Xu, et al., “Intelligent multicolor nano-sensor based on nontoxic dual fluoroprobe and MOFs for colorful consecutive detection of Hg2+ and cysteine,” J. Hazard. Mater 430, 128478 (2022). [CrossRef]
23. T. Chen, Z.Q. Dang, Z.M. Ding, et al., “Multibit NOT logic gate enabled by a function programmable optical waveguide,” Opt. Lett. 47(14), 3519–3522 (2022). [CrossRef]
24. R. Cheng, D.M. Zhang, J.H. Wang, et al., “Fluorinated photopolymer cascaded MMI based integrated optical waveguide switching matrix with encoding function,” Opt. Express 27(9), 12883–12898 (2019). [CrossRef]
25. J. Yue, X.Y. Sun, C.X. Wang, et al., “Triple-layered optical interconnecting integrated waveguide chip based on epoxy cross-linking fluorinated polymer photonic platform,” Opt. Express 31(12), 19415–19427 (2023). [CrossRef]
26. Q.M. Zhang, H.Y. Yu, M. Barbiero, et al., “Artificial neural networks enabled by nanophotonics,” Light: Sci. Appl. 8(1), 42 (2019). [CrossRef]
27. Y.P. Bai, X.Y. Xu, M.X. Tan, et al., “Photonic multiplexing techniques for neuromorphic computing,” Nanophotonics 12(5), 795–817 (2023). [CrossRef]
28. B.W. Bai, H.W. Shu, X.J. Wang, et al., “Towards silicon photonic neural networks for artificial intelligence,” Sci. China Inf. Sci. 63(6), 160403 (2020). [CrossRef]
29. L.J. Lu, S.Y. Zhao, L.J. Zhou, et al., “16 × 16 non-blocking silicon optical switch based on electro-optic Mach-Zehnder interferometers,” Opt. Express 24(9), 9295–9307 (2016). [CrossRef]
30. K. Suzuki, R. Konoike, S. Suda, et al., “Low-Loss, Low-Crosstalk, and Large-Scale Optical Switch Based on Silicon Photonics,” J. Lightwave Technol. 38(2), 233–239 (2020). [CrossRef]
31. K. Suzuki, R. Konoike, G.W. Cong, et al., “Strictly Non-Blocking 8 × 8 Silicon Photonics Switch Operating in the O-Band,” J. Lightwave Technol. 39(4), 1096–1101 (2021). [CrossRef]
