A DA-type narrow bandgap conjugated polymer, its preparation method, and opto-memristor devices based thereon and their applications.

Abstract

This application discloses a D-A type narrow bandgap conjugated polymer, its preparation method, and opto-memristor devices and applications based on it, belonging to the field of machine vision hardware technology. The absorption spectrum of the polymer covers the ultraviolet to near-infrared region. The heterojunction opto-memristor device constructed based on this polymer and zinc oxide exhibits photoresponse in a wide wavelength range from 365 nm to 1200 nm, and bidirectional reversible modulation of synaptic plasticity, both enhancing and inhibiting, can be achieved by controlling the wavelength and intensity of the light. The conductivity update of this device under light pulse modulation exhibits excellent linearity and high symmetry. A reservoir computing system constructed using it as a physical reservoir achieved an average recognition accuracy of 83% in fingerprint recognition tasks. This application provides a new material platform and device foundation for realizing broadband response, all-optical modulation neuromorphic visual perception.

Description

Technical Field

[0001] This application relates to the field of machine vision technology, and in particular to a DA-type narrow bandgap conjugated polymer, its preparation method, and opto-memristor devices and applications based thereon. Background Technology

[0002] With the rapid development of information technology, machine vision has become increasingly important as a core window for IoT systems to perceive and understand the external world. Biological vision systems, with their efficient parallel processing and adaptive perception capabilities, provide important insights for the development of next-generation machine vision. As a key carrier for realizing the integration of visual information sensing, storage, and computation, photoresistors can directly convert light signals into controllable electrical conductance states, making them one of the ideal platforms for constructing biomimetic visual synapses. Currently, various materials, including metal oxides, low-dimensional materials, perovskites, organic semiconductors, and biomaterials, are used to construct photoresistors. However, most existing devices have limited light response ranges, typically covering only visible light or a single wavelength band, and most can only achieve unidirectional (enhanced or suppressed) synaptic plasticity, making it difficult to fully simulate the complex information preprocessing and encoding functions in the biological retina.

[0003] Achieving broadband photoresponse is crucial for expanding the perceptual boundaries of artificial vision systems. The light information in nature far exceeds the range of human visible light. Broadening the response spectrum of devices to detect invisible light, such as ultraviolet and near-infrared light, will greatly enhance the environmental adaptability and information acquisition dimensions of machine vision. Simultaneously, simulating the bidirectional response of bipolar cells to light signals in the biological retina is essential. The coexisting excitatory (EPSC) and inhibitory (IPSC) postsynaptic currents constitute the neural basis for parallel processing of visual signals and contrast enhancement. Therefore, developing artificial photosynaptic devices with both broadband response and bidirectional photocurrent response capabilities is a key step towards efficient, intelligent, and retinal-like perception. To date, several broadband response or all-optical modulation material systems and corresponding memory mechanisms have been proposed. For example, broadband response can be achieved by constructing narrow bandgap materials using strategies such as two-dimensional van der Waals heterostructures (vdWHs) and quantum dots; all-optical modulation can be achieved by integrating positive and negative photoconductive materials in the same device to construct heterojunction structures and by controlling defects. Despite significant progress in optoelectronic memristors in recent years, achieving broadband response and all-optical modulation characteristics remains a major challenge due to limitations in device structure and physical models. Summary of the Invention

[0004] This application aims to address the technical challenges of existing optoelectronic memristor materials, such as limited optical response range and difficulty in simulating bidirectional synaptic plasticity (i.e., simultaneously achieving excitatory and inhibitory postsynaptic currents) in biological visual systems. This provides a new material platform and device foundation for constructing a high-efficiency intelligent biomimetic visual system with broadband response and all-optical modulation. The technical solution adopted in this application is as follows: The first aspect of this application provides a DA-type narrow bandgap conjugated polymer, which is polymerized from monomers M1 and M2. The structural formula of the monomer M1 is as follows: ; The structural formula of the monomer M2 is as follows: , Wherein, R1 is a C4-C20 straight-chain or branched alkyl, alkenyl or alkynyl group, wherein the alkyl group is optionally substituted with a halogen, alkoxy or cyano group; R2 is a C4-C30 straight-chain or branched alkyl, cycloalkyl, aryl or aralkyl group; M is SnBu3 or SnMe3; Y is a halogen or sulfonate ester.

[0005] In this application, the DA-type narrow bandgap conjugated polymer includes a donor unit and an acceptor unit, wherein the monomer M1 provides the donor unit and the monomer M2 provides the acceptor unit.

[0006] In some embodiments of this application, R1 is a C4-C18 straight-chain alkyl group. In some embodiments of this application, R2 is a C3-C20 branched alkyl group.

[0007] In some specific embodiments of this application, the structural formula of the DA-type narrow bandgap conjugated polymer is as follows: , Where n is an integer, and the range is n≧2. In this application, there is no special requirement for the upper limit of the polymer size; for example, it can be 5000.

[0008] The second aspect of this application provides a method for preparing the DA-type narrow bandgap conjugated polymer described in the first aspect of this application, comprising the following steps: carrying out the Stille coupling reaction of the monomer M1 and the monomer M2 in the presence of a palladium catalyst and a ligand to obtain the DA-type narrow bandgap conjugated polymer.

[0009] Stille coupling reaction refers to the C-C bond coupling reaction between organotin reagents and halogenated or halogen-like products in the presence of a palladium catalyst.

[0010] In some embodiments of this application, the palladium catalyst is Pd2(dba)3; the ligand is P(o-tol)3.

[0011] In some embodiments of this application, the Stille coupling reaction is carried out in an organic solvent at a temperature of 110-130°C for 24-72 hours. In some specific embodiments of this application, the organic solvent is toluene.

[0012] In some embodiments of this application, the Stille coupling reaction is carried out in an inert gas because oxygen oxidizes the palladium catalyst and causes the organotin compound to self-couple.

[0013] In some embodiments of this application, the monomer M1 is prepared using the following steps:

[0014] The specific steps are as follows: Will and An etherification reaction was carried out in a first organic solvent in the presence of a first catalyst to obtain... ; Will and The halogenation reaction was carried out in a second organic solvent to obtain ; Will The Yamamoto coupling reaction was carried out in a third organic solvent in the presence of a second catalyst to obtain... ; Will MX3 undergoes a lithiation-tinning reaction with a third catalyst in a fourth organic solvent to obtain the monomer M1. The third catalyst is a lithium catalyst, and X1 and X3 are each independently selected from one of the halogens. Each of the above steps is carried out in an inert gas.

[0015] In some embodiments of this application, the first catalyst is toluenesulfonic acid monohydrate, which catalyzes... The methoxy group is replaced by R1 (acid-catalyzed ether exchange reaction), thus yielding .

[0016] In some embodiments of this application, the second catalyst is a stable zero-valent nickel complex with a specific coordination environment generated in situ using Ni(COD)2, 2,2'-bipyridine (bpy) and extracyclooctadiene (COD), which can catalyze reactions such as cross-coupling, olefin polymerization or functional group transformation.

[0017] In some embodiments of this application, the third catalyst is lithium diisopropylamide (LDA), which is prepared by reacting anhydrous diisopropylamine with n-butyllithium under low temperature conditions. As a sterically hindered non-nucleophilic strong base (pKa=35.7), it can deprotonate carbonyl compounds containing α-hydrogen and can be used to catalyze reactions such as Aldol condensation, enolization, and α-alkylation of 1,3-dicarbonyl compounds.

[0018] In some embodiments of this application, the first organic solvent, the second organic solvent, the third organic solvent, and the fourth organic solvent are each independently selected from one or more of acetonitrile, DMSO, DMF, DCM, acetone, DEMA, HMPA, triethylamine, CH2Cl2, pyridine, chloroform, chlorobenzene, THF, dioxane, diethyl ether, benzene, and toluene.

[0019] In some embodiments of this application, the monomer M2 is prepared using the following steps:

[0020] The specific steps are as follows: Will and Ullmann-type amination was carried out in a fifth organic solvent in the presence of a fourth catalyst to obtain ; Will and An acylation cyclization reaction was carried out in a sixth organic solvent to obtain... ; Will and The Knoevenagel condensation reaction was carried out in the seventh organic solvent to obtain ; Each of the above steps is carried out in an inert gas. Will and The substitution reaction was carried out in an eighth organic solvent to give the monomer M2. In this context, X4 represents a halogen, and Y represents a halogen or a sulfonate.

[0021] In some embodiments of this application, the fifth catalyst is a copper catalyst. Preferably, the copper catalyst comprises copper powder and cuprous iodide (CuI), wherein CuI is one of the most commonly used copper catalysts, providing monovalent copper (Cu). +CuI is the active catalytic center of CN coupling. Copper powder (Cu) can act as an auxiliary reducing agent or a co-catalyst to help maintain the appropriate valence state of copper, or to inhibit catalyst deactivation by "sacrificing copper". In some systems, copper powder can also promote the dispersion and recycling of copper catalysts. The combined use of CuI and copper powder can sometimes improve reaction efficiency, especially in the presence of ligands to form more active catalytic species.

[0022] In some embodiments of this application, and The reaction system also includes tripotassium phosphate (K3PO4), which deprotonates the amine to generate a more nucleophilic amine anion or enhances the coordination and attack capabilities of the amine; at the same time, it neutralizes the HX2 produced in the reaction, pushing the reaction equilibrium to the right. K3PO4 is a strong but non-nucleophilic base and is commonly used in CN coupling.

[0023] In some embodiments of this application, the fifth, sixth, seventh, and eighth organic solvents are each individually selected from one or more of acetonitrile, DMSO, DMF, DCM, acetone, DEMA, HMPA, triethylamine, CH2Cl2, pyridine, chloroform, chlorobenzene, THF, dioxane, diethyl ether, benzene, and toluene.

[0024] In some embodiments of this application, the reaction formulas for monomers M1 and M2 are as follows: the Stille coupling reaction is carried out in a ninth solvent in the presence of a palladium catalyst.

[0025]

[0026] In some embodiments of this application, the ninth organic solvent is selected from one or more of acetonitrile, DMSO, DMF, DCM, acetone, DEMA, HMPA, triethylamine, CH2Cl2, pyridine, chloroform, chlorobenzene, THF, dioxane, diethyl ether, benzene, and toluene.

[0027] A third aspect of this application provides an opto-memristor device, comprising: First electrode; Second electrode; and An active layer is disposed between the first electrode and the second electrode; The active layer comprises a polymer layer formed from a DA-type narrow bandgap conjugated polymer as described in any of the first aspects of this application.

[0028] In some embodiments of this application, the active layer further includes a metal oxide layer, and the polymer layer and the metal oxide layer form a heterojunction structure.

[0029] In some specific embodiments of this application, the metal oxide layer is a zinc oxide layer.

[0030] In some embodiments of this application, the opto-memristor device has a structure in which a first electrode / polymer layer / metal oxide layer / second electrode are stacked in sequence.

[0031] In some specific embodiments of this application, the first electrode is indium tin oxide and the second electrode is indium tin oxide.

[0032] The fourth aspect of this application provides the application of any of the opto-memristor devices described in the third aspect of this application in constructing a neuromorphic computing system, wherein the neuromorphic computing system is a reservoir computing system used for image recognition.

[0033] In some embodiments of this application, the image recognition is fingerprint recognition.

[0034] In some embodiments of this application, the readout layer weight update in the reservoir computing system is achieved using a fully optical modulation method, wherein weight enhancement is achieved by applying infrared light pulses, and weight suppression is achieved by applying ultraviolet light pulses.

[0035] Compared with the prior art, this application has the following advantages: 1. The DA-type narrow bandgap conjugated polymer PBTP of this application forms a highly coplanar main chain conformation and excellent intramolecular charge transfer characteristics through the conjugation of donor and acceptor units. Its absorption spectrum covers the ultraviolet, visible, and near-infrared regions, exhibiting particularly strong absorption in the second near-infrared window (990 nm), providing a key material basis for achieving broadband photoelectric responses in devices. Simultaneously, this polymer exhibits good solubility in common organic solvents and a high thermal decomposition temperature of 388℃, demonstrating excellent processing performance and thermal stability, meeting the requirements for device fabrication and application.

[0036] 2. The opto-memristor device of this application forms an ITO / PBTP / ZnO / ITO stacked structure by compositing a PBTP polymer layer with a ZnO metal oxide layer. This heterojunction interface effectively promotes the separation and transport of photogenerated carriers. Combined with the broad-spectrum absorption characteristics of PBTP, the device exhibits a stable photoresponse across a wide wavelength range of 365 nm to 1200 nm. More importantly, by applying infrared and ultraviolet light pulses respectively, the device can simulate the plasticity of biological synapses, achieving bidirectional reversible modulation of both enhancing and inhibiting postsynaptic currents, overcoming the limitation of existing devices that can only achieve unidirectional responses.

[0037] 3. The opto-memristor device of this application exhibits excellent linearity in conductivity updates under optical pulse modulation (nonlinearity factor α).p =0.05, α d It exhibits high symmetry (0.12) and high asymmetry (0.07), and demonstrates highly consistent performance across different devices, possessing good repeatability and operational reliability. This lays a solid foundation for building high-precision, low-power hardware neural networks.

[0038] 4. The photoelectric memristor device of this application can be used as the physical reservoir of the reservoir computing system. Utilizing its rich nonlinear dynamic characteristics, the encoded fingerprint image light pulse sequence is mapped into high-dimensional separable features. Combined with the readout layer training of full optical modulation, an average recognition accuracy of 83% was finally achieved in the fingerprint recognition task.

[0039] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0040] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of this application are illustrated in the drawings by way of example and not limitation, in which: Figure 1 The thermogravimetric analysis curve of polymer PBTP in Example 1 of this application is shown; Figure 2 The ultraviolet-visible-near-infrared absorption spectra of PBTP toluene solution and PBTP film in Example 1 of this application are shown. Figure 3 The following are the DFT calculation results of the simplified side chains of the PBTP molecular structure in Example 1 of this application at the B3LYP / 6-31G(d) level: (a) Molecular geometry calculated by density functional theory at the B3LYP / 6-31G(d) theoretical level; (b) Frontier orbital density distribution diagram of PBTP; (c) Schematic diagram of hydrogen bonds within the PBPT molecule, with dashed lines representing hydrogen bonds and bond angles / bond lengths being schematic. Figure 4 A schematic diagram of the ITO / PBTP / ZnO / ITO device structure in Embodiment 2 of this application is shown; Figure 5 The cross-sectional SEM image of ITO / PBTP / ZnO in Embodiment 2 of this application is shown; Figure 6 The AFM images of the PBTP film, ZnO film, and PBTP / ZnO film in Example 2 of this application are shown, with a scan size of 5×5μm. 2 ; Figure 7This illustrates the synaptic behavior of the ITO / PBTP / ZnO / ITO device in Embodiment 2 of this application under stimulation by six consecutive light pulses of different wavelengths at a read voltage of 0.1V (light intensity: 50mW / cm²). 2 ); Figure 8 This illustrates the synaptic behavior of the ITO / PBTP / ZnO / ITO device under weak ultraviolet light pulse stimulation at a read voltage of 0.1V in Embodiment 2 of this application (light intensity: 10mW / cm²). 2 ); Figure 9 The IV curves of the ITO / PBTP / ZnO / ITO device under different wavelengths of light at a read voltage of 0.1V are shown in Embodiment 2 of this application. Figure 10 The following are examples from Embodiment 3 of this application: (a) a schematic diagram of a reservoir computing system architecture for fingerprint recognition; (b) LTP and LTD curve fitting of an ITO / PBTP / ZnO / ITO photosynaptic device, wherein the enhancement process is excited by light at a wavelength of 1200 nm and the suppression process is excited by light at a wavelength of 365 nm; (c) current measurements on 20 randomly selected devices during the enhancement and suppression processes; (d) statistical distribution of device current during the enhancement process; and (e) statistical distribution of device current during the suppression process. Figure 11 The reservoir calculation sampling method in Embodiment 3 of this application is shown; Figure 12 The following are examples from Embodiment 3 of this application: (a) the EPSC response of the ITO / PBTP / ZnO / ITO device to 16 different pulse currents with a readout voltage of 0.1V; (b) the statistical results of 16 states measured on 20 randomly selected devices, including box plots and normal distribution descriptions; (c) a schematic diagram of fingerprint image preprocessing, reservoir encoding, and output, where the pattern is converted into a pulse current as input to the readout layer; (d) the accuracy variation of fingerprint dataset recognition using the device (simulation); (e) the loss curve during training; and (f) the confusion matrix based on the system. Detailed Implementation

[0041] Unless otherwise stated, implied from the context, or as is customary in the art, all testing and characterization methods used in this application are concurrent with the filing date of this application. Where applicable, any patent, patent application, or disclosure relating to this application is incorporated herein by reference in its entirety, and its equivalent patent families are also incorporated herein by reference, in particular the definitions of relevant terms in the art disclosed in such documents. If any definition of a specific term disclosed in the prior art is inconsistent with any definition provided in this application, the definition provided in this application shall prevail.

[0042] To make the technical problems, technical solutions and beneficial effects solved by this application clearer, the following detailed description is provided in conjunction with embodiments.

[0043] The following examples are used to illustrate preferred embodiments of this application. Those skilled in the art will understand that the techniques disclosed in the examples represent technologies discovered by the inventors that can be used to implement this application, and therefore can be considered preferred embodiments of this application. However, those skilled in the art should understand from this specification that many modifications can be made to the specific embodiments disclosed herein, still yielding the same or similar results, without departing from the spirit or scope of this application.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains, and all materials cited herein and referenced by them are incorporated herein by reference.

[0045] Those skilled in the art will recognize, or can learn through routine experimentation, many equivalents of specific embodiments of the invention described herein. These equivalents will be included in the claims.

[0046] Unless otherwise specified, the experimental methods used in the following examples are conventional methods. Unless otherwise specified, the instruments and equipment used in the following examples are all conventional laboratory instruments and equipment; unless otherwise specified, the experimental materials used in the following examples were all purchased from conventional biochemical reagent stores.

[0047] Example 1: Preparation of DA-type narrow bandgap conjugated polymers 1. Synthesis of monomer M1 The synthetic route for monomer M1 is shown below:

[0048] Specifically, the synthesis steps are as follows: (1) Synthesis of 3-(undecyloxy)thiophene (compound 1) Under an argon atmosphere, 3-methoxythiophene (5.00 g, 43.80 mmol), undecanol (15.2 g, 87.6 mmol), p-toluenesulfonic acid monohydrate (p-TsOH) (0.84 g, 0.05 equivalent), and 40 mL of toluene were added sequentially to a 250 mL Schlenk flask. The reaction mixture was heated to 110 °C and stirred under reflux overnight. After the reaction was complete, the mixture was cooled to room temperature and extracted with dichloromethane (DCM). The organic phase was collected, washed with water, and dried over anhydrous magnesium sulfate. The drying agent was removed by filtration, and the filtrate was concentrated on a rotary evaporator to remove the solvent. The crude product was purified by column chromatography using petroleum ether as the eluent to give 6.00 g of a white solid, with a yield of 54%, which is compound 1. The 1H NMR spectrum data are as follows: 1 H NMR(400MHz, CDCl3, ppm): δ 7.19(dd, 1H), 6.78(dd, 1H), 6.25(dd,1H), 3.99(t, 2H), 1.85-1.71(m, 2H), 1.53-1.14(m, 16H), 0.97(t, 3H). (2) Synthesis of 2-bromo-3-(undecyloxy)thiophene (compound 2) Under argon protection, compound 1 (2.5 g, or 9.8 mmol) was dissolved in 20 mL of N,N-dimethylformamide (DMF) and placed in a reaction flask. N-bromosuccinimide (NBS) (1.83 g, or 10.29 mmol) was added to the system with stirring. The reaction mixture was first stirred at 0 °C for 30 min, then allowed to rise naturally to room temperature for 12 h. After the reaction was complete, water was added to quench the reaction, and the mixture was extracted three times with ethyl acetate (EA). The combined organic phases were dried over anhydrous MgSO4 and concentrated under reduced pressure to remove the solvent. The crude product was purified by column chromatography using petroleum ether as eluent, yielding 2.8 g of a white solid product (85% yield), which was compound 2. The 1H NMR spectrum data are as follows: 1 H NMR (400MHz, CDCl3, ppm): δ 7.21(d, 1H), 6.78(d, 1H), 4.08(t, 2H), 1.81-1.69(m, 2H), 1.40-1.20(m, 16H), 0.93(t, 3H). (3) Synthesis of 3,3'-bis(undecyloxy)-2,2'-bithiophene (compound 3) Under an argon atmosphere, Ni(COD)₂ (1.75 g, 5 mmol), 2,2′-bipyridine (bpy) (810 mg, 5 mmol), and cyclooctadiene (COD) (0.9 mL, 5 mmol) were added to a Schlenk flask, followed by the addition of 20 mL of ultra-dry DMF. The mixture was heated to 55 °C and stirred for 30 min to activate the catalyst. Subsequently, compound 2 (460 mg, 1.4 mmol) was dissolved in 40 mL of ultra-dry toluene, and this solution was added to the reaction system using a syringe. The mixture was heated to 80 °C and stirred continuously for 3 days. After the reaction was complete, the mixture was cooled to room temperature, and toluene was removed by rotary evaporation under reduced pressure. 50 mL of DCM was added to the residue to dissolve the product, and the mixture was filtered to remove insoluble impurities. The filtrate was washed three times with water, dried over anhydrous MgSO₄, and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography using a mixed solvent of DCM:petroleum ether = 1:3 as the eluent, yielding 200 mg of a yellow solid product, with a yield of 43%, which is compound 3. The 1H NMR spectral data are as follows: 1 H NMR(400MHz, CDCl3, ppm): δ 7.11(d, 2H), 6.87(d, 2H), 4.12(t, 4H),1.90-1.78(m, 4H), 1.55-1.49(m, 4H), 1.28-1.23(m, 28H), 0.91(t, 6H). (4) Synthesis of (3,3'-bis(undecyloxy)-[2,2'-bithiophene]-5,5'-diyl)bis(trimethylstanane) (monomer M1) Under argon protection, compound 3 (300 mg, or 0.56 mmol) was placed in a Schlenk flask and dissolved in 10 mL of ultra-dry tetrahydrofuran (THF). The reaction flask was cooled to -78 °C, and then diisopropylaminolithium (LDA, n-BuLi) (2.0 M THF solution, 2.4 mL, or 4.80 mmol) was slowly added dropwise. After the addition was complete, the mixture was stirred at -78 °C for 1 h. SnMe3Cl (1.0 M THF solution, 4.8 mL, or 4.8 mmol) was then added to the system. After the addition was complete, the reaction mixture was allowed to rise naturally to room temperature, and stirring was continued for 12 h. After the reaction was complete, the mixture was quenched with water and extracted with petroleum ether (PE). The combined organic phases were washed successively with water and saturated brine, and dried over anhydrous MgSO4. After filtration, the solvent was removed by concentration under reduced pressure to obtain a grayish-green solid crude product. The crude product was purified by recrystallization from DCM / methanol to obtain 150 mg of a light green solid, with a yield of 32%, which is the monomer M1. The 1H NMR data are as follows: 1H NMR(400MHz, CDCl3, ppm): δ 6.92(s, 2H), 4.15(m, 4H), 1.91(m, 4H), 1.37-1.14(m, 32H), 0.96-0.76(t, 8H), 0.76-0.25(m, 16H). 2. Synthesis of monomer M2 The synthetic route for monomer M2 is shown below:

[0049] Specifically, the synthesis steps are as follows: (1) Synthesis of N-(2-octyldodecyl)thiophene-3-amine (compound 4) and 4-(2-octyldodecyl)-4H-thiopheno[3,2-b]pyrrole-5,6-dione (compound 5) Under argon protection, 3-bromothiophene (8.97 g, 55 mmol), copper powder (0.18 mg, 2.75 mmol), cuprous iodide (CuI) (0.52 g, 2.75 mmol), tripotassium phosphate (K3PO4) (2.92 g, 137.5 mmol), 2-octyldodecaneamine (24.55 g, 82.5 mmol), and N,N-dimethylethanolamine (DEMA) (15 mL) were added to a reaction flask. The mixture was heated to 80 °C and stirred at that temperature for 48 h. After the reaction was completed, the mixture was cooled to room temperature, and insoluble matter was removed by filtration. The filter cake was washed with polyethylene (PE). The filtrates were combined and concentrated by rotary evaporation. The crude product was purified by silica gel column chromatography using petroleum ether:PE = 5:1 as the eluent to give 5.02 g of a brown oily product, with a yield of 42%, which is compound 4.

[0050] Under an argon atmosphere, oxalyl chloride (4.76 g, 37.5 mmol) was dissolved in anhydrous DCM (20 mL) and cooled to 0 °C. Compound 4 (9.48 g, 25 mmol) in anhydrous DCM solution (20 mL) was slowly added dropwise over 30 min. After stirring for another 30 min, triethylamine (Et3N) (7.6 g, 75 mmol) in anhydrous DCM solution (20 mL) was slowly added dropwise over another 30 min. After the addition was complete, the ice bath was removed, and the mixture was allowed to rise naturally to room temperature with stirring for 12 h. The reaction was quenched with water, and the mixture was extracted three times with DCM. The organic phases were combined, dried over anhydrous MgSO4, and concentrated under reduced pressure. The residue was purified by silica gel column chromatography using petroleum ether:PE = 5:1 as the eluent to give 3.80 g of compound 5, with a yield of 35%. The 1H NMR spectrum data are as follows: 1H NMR (400MHz, CDCl3, ppm): δ 7.98(s, 1H), 6.75(s, 1H), 3.54(d, 2H), 1.77(m, 1H), 1.28(m, 32H), 0.86(m, 6H). (2) Synthesis of (3E,6E)-3,6-bis(4-(2-octyldodecyl)-5-oxo-4,5-dihydro-6H-thieno[3,2-b]pyrrole-6-ylidene)piperazine-2,5-dione (compound 6) At room temperature, 1,4-diacetylpiperazine-2,5-dione (515 mg, or 2.6 mmol), compound 5 (2.5 g, or 5.78 mmol), chloroform (CHCl3) (15 mL), DMF (5 mL), and triethylamine (0.8 mL, or 5.72 mmol) were added sequentially to a 150 mL round-bottom flask. The reaction mixture was stirred for 6 h. After the reaction was complete, the mixture was slowly added dropwise to methanol, precipitating out a solid product which was collected by vacuum filtration. The crude product was purified by silica gel column chromatography using petroleum ether:DCM = 3:1 as the eluent to give 0.96 g of compound 6, with a yield of 35%. The 1H NMR spectrum data are as follows: 1 H NMR(400MHz, CDCl3, ppm): δ 12.14(s, 2H), 7.47(s, 2H), 6.69(s, 2H), 3.64 (d, 4H), 1.84(m, 2H), 1.29 (m, 34H), 0.85(m,12H). (3) Synthesis of (3E,6E)-3,6-bis(2-bromo-4-(2-octyldodecyl)-5-oxo-4,5-dihydro-6H-thieno[3,2-b]pyrrole-6-ylidene)piperazine-2,5-dione (monomer M2) Compound 6 (0.35 g, or 0.33 mmol) was dissolved in 20 mL of THF, and NBS (0.12 g, or 0.69 mmol) was added with stirring. The reaction system was first stirred at 0 °C for 15 min, then allowed to rise naturally to room temperature and stirred overnight. After the reaction was complete, water was added to quench the reaction, and the mixture was extracted three times with DCM. The organic phases were combined, dried with anhydrous desiccant, and concentrated under reduced pressure to remove the solvent. The crude product was purified by silica gel column chromatography using petroleum ether:DCM = 3:1 as the eluent to give monomer M2, with a mass of 0.53 g and a yield of 36%. The 1H NMR data are as follows: 1H NMR(400MHz, CDCl3, ppm): δ 12.03(s, 2H), 6.84(s, 2H), 3.58(d, 4H), 1.86(m, 2H), 1.19(m, 34H), 0.88(m,12H). 3. Synthesis of the conjugated polymer PBTP The synthetic route for the conjugated polymer PBTP is shown below:

[0051] Specifically, the synthesis steps are as follows: Under an argon atmosphere, monomers M1 (83.2 mg, 0.01 mmol), M2 (110.2 mg, 0.01 mmol), Pd2(dba)3 (5 mg, 0.005 mmol), P(o-tol)3 (6.7 mg, 0.02 mmol), and 20 mL of toluene were added sequentially to a Schlenk tube. The system was sealed and deoxygenated by a three-cycle liquid nitrogen freezing-vacuuming-thawing process. The reaction tube was then placed in a 120°C oil bath and stirred for 48 h. After the reaction was complete and cooled to room temperature, the reaction solution was slowly added dropwise to methanol to precipitate. The precipitated crude solid product was collected by filtration. This crude product was purified by Soxhlet extraction using THF as the solvent. The extract was concentrated under reduced pressure and then precipitated again in methanol. The precipitate was collected by filtration, dried under vacuum, and yielded 129 mg of a black solid product, with a yield of 90%.

[0052] The polymer was determined using a Waters 2690 gel permeation chromatography (GPC) system, with polystyrene as the standard and tetrahydrofuran as the mobile phase. The polymer had a number-average molecular weight (Mn) of 10754, a weight-average molecular weight (Mw) of 21692, a peak molecular weight (Mp) of 14556, a mass-average molecular weight (Mz) of 38801, and a polymer dispersion index (PDI) of 2.02.

[0053] The proton NMR data are as follows: 1 H NMR(400MHz, CDCl3, ppm): δ 8.30-7.93(m), 4.86-4.36(m), 4.35-3.49(m). 2.14-1.68(m), 1.59-1.46(m), 1.25-1.43(m), 0.75-0.96(m). 4. PBTP Performance Testing Experiments have shown that the polymer PBTP exhibits good solubility in common organic solvents such as toluene, THF, and DMF.

[0054] The inventors characterized the thermal stability of the synthesized polymer using a TGA-Q500 thermogravimetric analyzer. The thermogravimetric analysis (TGA) curves are shown below. Figure 1 As shown, PBTP exhibits good thermal stability, with an initial decomposition temperature of 388℃, indicating that the material can meet the temperature requirements for device operation.

[0055] The inventors further used a Shimadzu UV-3600 Plus spectrophotometer to measure the UV-Vis-NIR absorption spectrum of PBTP. The UV-Vis-NIR absorption spectra of PBTP in toluene solution and spin-coated thin film states are as follows: Figure 2 As shown. By Figure 2 It is known that PBTP exhibits a broad absorption spectrum in the second window of the near-infrared spectrum, with a weak absorption peak at 449 nm, attributed to the π–π* transition; its maximum absorption peak is located at 990 nm, corresponding to the intramolecular charge transfer process. The absorption range of PBTP covers the ultraviolet, visible, and near-infrared regions.

[0056] The inventors performed DFT calculations on the simplified side chains at the B3LYP / 6-31G(d) level of the PBTP molecular structure, such as... Figure 3 In the middle a, the optimized geometry of PBTP exhibits a highly coplanar main chain conformation, with dihedral angles of 1.08° and 0.26° between neighboring units, indicating good charge transport properties. This is because the intramolecular hydrogen bonds of the acceptor unit contribute to the formation of a rigid and planar main chain. Figure 3 (c). Additionally, such as... Figure 3 As shown in Figure b, the electron cloud of the highest occupied molecular orbital (HOMO) in PBTP is mainly distributed on the 3,3'-bis(undecyloxy)-2,2'-bithiophene segment, while the electron cloud of the lowest unoccupied molecular orbital (LUMO) is mainly distributed on the (3E,6E)-3,6-bis(4-(2-octyldodecyl)-5-oxo-4,5-dihydro-6H-thiopheno[3,2-b]pyrrole-6-ylidene)piperazine-2,5-dione segment. These results indicate the existence of conjugated transport channels for charge carriers in PBTP. The asymmetric electron cloud distribution of the HOMO and LUMO orbital energy levels suggests the presence of charge transfer processes within PBTP.

[0057] Example 2: Fabrication and Performance Testing of Opto-memristor Devices 25 µL of a toluene solution (2.5 mg / mL) of PBTP prepared in Example 1 was spin-coated onto an indium tin oxide (ITO) glass substrate to form a film: first spin-coated at 800 rpm for 15 s, then at 2000 rpm for 55 s. The film was vacuum-dried overnight at 60 °C to remove residual solvent. Subsequently, a ZnO film was deposited on the PBTP film (PBTP layer) by magnetron sputtering (using a ZnO ceramic target, under an Ar atmosphere, RF power 80–120 W). Finally, a patterned ITO top electrode (50 nm thick) was deposited by magnetron sputtering using a mask, thereby obtaining an opto-memristor device with an ITO / PBTP / ZnO / ITO structure, wherein the PBTP / ZnO heterostructure serves as the active layer. The device structure is shown in the figure. Figure 4 As shown. All electrical measurements were performed under ambient conditions using a Keithley 4200 semiconductor parameter analyzer without any device packaging.

[0058] Figure 5 The image shows a cross-sectional SEM image of the ITO / PBTP / ZnO layer, clearly distinguishing each layer. The ITO bottom electrode is approximately 200 nm thick, the PBTP layer is approximately 20 nm thick, and the ZnO film deposited by magnetron sputtering is approximately 16 nm thick. The surfaces of each film are smooth, and the interlayer interfaces are clear and tightly bonded, which is conducive to the efficient transport of charge carriers between layers.

[0059] To further characterize the surface morphology of the films, the inventors used AFM to test ZnO, PBTP monolayer films, and PBTP / ZnO bilayer films. The results are as follows: Figure 6 As shown, the average surface roughness of the ZnO film was measured to be 2.41 nm, the average surface roughness of the PBTP film was 1.95 nm, and the average surface roughness of the PBTP / ZnO bilayer film was 2.64 nm. Overall, the surfaces of all films are relatively smooth, which is beneficial to the stability of device performance.

[0060] At a readout voltage of 0.1V, the inventors systematically investigated the photoresponse behavior of the ITO / PBTP / ZnO / ITO device to light pulses of different wavelengths. Benefiting from the broad-spectrum absorption characteristics of PBTP, the device exhibited a significant photoresponse in the wavelength range of 365nm to 1200nm. First, the inventors applied five consecutive light pulses with wavelengths of 365, 430, 560, 650, 850, and 1200nm (light intensity: 50mW / cm²). 2 (Pulse width: 2s, interval: 3s), results are as follows Figure 7 As shown, under stimulation by light pulses of different wavelengths, the device current exhibits a successively increasing response trend. Subsequently, a light pulse with a wavelength of 365 nm and an intensity of 10 mW / cm² was used. 2The device was stimulated by continuous weak ultraviolet light pulses, and the device current was observed to decrease successively under multiple pulses. Figure 8 The above results demonstrate that by adjusting the wavelength and intensity of light, the device can achieve stable bidirectional synaptic plasticity, enabling reversible control of both enhancement and suppression modes. Furthermore, the IV characteristic curves of the device under different wavelengths of light are shown below. Figure 9 As shown, the device exhibits rectification characteristics, indicating that a Schottky contact has been formed between the semiconductor layer and the electrode.

[0061] Example 3: Memristor-based reservoir computing system for fingerprint recognition Reservoir computing (RC) is a highly energy-efficient artificial intelligence computing paradigm inspired by biological neural networks. Its core structure consists of three parts: an input layer, a reservoir, and a readout layer. The reservoir itself is a fixed, high-dimensional, and complex dynamic nonlinear system. Its role is to nonlinearly map low-dimensional input signals to a high-dimensional feature space, requiring only one layer of connection weights between the reservoir and the readout layer to be trained. In contrast, traditional deep neural networks (DNNs) require training a massive number of synaptic connections across all layers of the network and performing intensive vector-matrix multiplication operations during inference, leading to a significant increase in computational complexity and energy consumption. Therefore, the RC architecture can greatly simplify learning algorithms, reduce training costs, and is particularly suitable for efficient information processing utilizing the inherent complex dynamics of physical devices.

[0062] Based on the above advantages, the inventors built a hardware RC system prototype using ITO / PBTP / ZnO / ITO devices and successfully applied it to image recognition tasks to verify its application potential in neuromorphic visual information processing.

[0063] In a reservoir computing system constructed using the prepared ITO / PBTP / ZnO / ITO opto-memristor devices, the inventors completed the task of fingerprint image recognition. Figure 10 Figure a demonstrates the complete process of fingerprint recognition implemented by the RC system: First, the raw fingerprint image obtained from the standard fingerprint database (FVC2002, DB1_B subset) is preprocessed and encoded into a series of light pulse sequences corresponding to pixel grayscale values. These light pulse sequences, as spatiotemporal input signals, are applied to a physical reservoir composed of multiple photoresistive devices. The rich nonlinear dynamics within the reservoir map the simple input pulses into high-dimensional, separable "reservoir states." These states are then passed to the readout layer. The readout layer consists of a set of adjustable memristor synapses, whose weights are trained using a linear regression algorithm, ultimately completing the mapping and recognition from the feature space to a specific fingerprint category.

[0064] During the online training of the readout layer, the synaptic weights are updated entirely through optical modulation. Specifically, the weight value (W) of each synapse is determined by the electrical conductance difference (G) between a pair of devices. + - G - The weight enhancement is achieved by applying an infrared (1200nm) pulse, which selectively increases G. + And reduce G - Weight suppression, on the other hand, is achieved by applying weak ultraviolet (365nm) pulses, which has the opposite effect. This all-optical modulation strategy avoids complex electrical writing operations and simplifies system control.

[0065] To quantitatively evaluate the performance of this memristor device in simulating synaptic weight updates, the inventors performed nonlinear fitting on its LTP and LTD processes under continuous light pulses. Specifically, the conductance Gn after the nth pulse is described by the following equation:

[0066] in G max and G min These represent the maximum and minimum achievable conductance values, respectively. β is the learning rate or update factor, and α represents the nonlinearity factor. A plus sign (+) indicates enhancement (LTP), and a minus sign (-) indicates suppression (LTD).

[0067] In addition, the nonlinear parameter α p and α d The conductivity changes during the enhancement and suppression phases were obtained by curve fitting. α=0 corresponds to ideal linear weight updates, while higher or lower values ​​represent sublinear or superlinear behavior, respectively. The asymmetry of the conductivity update can be defined by the following equation: Asymmetry=|α p -α d | Low alpha values ​​and low asymmetry are crucial for neuromorphic computing tasks because they help achieve uniform synaptic weight updates and improve the learning accuracy of hardware neural networks.

[0068] The extracted key parameters show that the device's conductance update has excellent linearity, with a nonlinearity factor α. p =0.05, α d =0.12, and its symmetry is very high (asymmetry |α p -α d |=0.07), such as Figure 10As shown in Figure b, low nonlinearity and high symmetry are key to achieving high-precision, low-power neuromorphic computing, indicating that this device has significant application potential in hardware neural networks.

[0069] Furthermore, to ensure the reliability of the system in actual integration, the inventors systematically tested its optically modulated LTP-LTD behavior on 20 independently fabricated devices. For example... Figure 10 As shown in Figure c, all devices exhibit highly consistent conductance change trajectories under the same light pulse. Furthermore, the inventors statistically analyzed the peak current distributions of these 20 devices after experiencing 20 infrared light pulses and 20 ultraviolet light pulses respectively. Figure 10 (d, e). Statistical results show that the conductivity changes are uniform among the devices and the current values ​​are concentrated, which fully demonstrates that the device fabrication process has good repeatability and that the device itself has excellent inter-device consistency and operational reliability.

[0070] To thoroughly evaluate the hardware feasibility of this ITO / PBTP / ZnO / ITO opto-memristor device as a physical reservoir in a reservoir computing system, the inventors first explored its capabilities in high-precision state mapping. Specifically, the inventors chose to encode 4-bit binary numbers (from "0000" to "1111") into a sequence of infrared light pulses with a spatiotemporal structure, aiming to verify whether the device could generate a sufficient number of stable high-dimensional conductance states that correspond one-to-one with the input pattern. The encoding process is as follows: Figure 11 As shown: Each binary bit corresponds to a 2-second time unit, where a logic "1" indicates that a 2-second infrared light pulse with a wavelength of 1200nm (intensity 50mW / cm²) is applied within that time period. 2 ), while logic "0" indicates that the device remains in a dark state during this period. Within an 8-second encoding cycle, the inventors measured the device current at five equal time intervals (0s, 2s, 4s, 6s, and 8s) under a constant readout voltage of 0.1V to monitor the evolution of the conductance state in real time. Experimental results show that at the end of the encoding cycle, 16 different input patterns were successfully mapped to 16 statistically distinct and well-separated steady-state conductance values ​​( ). Figure 12 (a) This provides a rich feature base for subsequent pattern classification.

[0071] To further verify the robustness and scalability of the encoding scheme to device performance, the inventors repeated the above encoding test on 20 independently fabricated devices. Figure 12As shown in Figure b, all devices exhibit highly consistent conductance response trajectories and final steady-state values ​​for the same binary input mode, and their small inter-device deviations fully demonstrate the excellent fabrication process uniformity and operational reliability. The inventors selected data from a subset DB1_B of the publicly available FVC 2002 fingerprint database, which contains 80 original fingerprint images from 10 different collectors. To enhance the model's generalization ability and expand the training data, the inventors performed data augmentation on each original image, including center cropping and offset cropping along eight different directions, ultimately generating an augmented dataset containing 720 images. Each fingerprint image was first binarized and resized, then reorganized column-wise: every four consecutive columns of data in the image matrix were rearranged into an 80×4 submatrix, where each row corresponds to a 4-bit binary vector. Following the aforementioned encoding process, each 4-bit vector was converted into an 8-second spatiotemporal optical pulse sequence, which was input into the physical reservoir composed of memristor devices. The reservoir generates a unique high-dimensional dynamic response to each pulse sequence, from which the inventors extract features. Ultimately, each fingerprint sample can generate 80 sets of high-dimensional feature vectors for subsequent training. Figure 12 (c)

[0072] The training of the readout layer was simulated in software. The inventors constructed a single-layer perceptron as the readout layer network, trained using feature vectors collected from the hardware storage layer. After 30 training epochs, the system achieved an average recognition accuracy of 83% on the test set. Figure 12 (d). The loss function value steadily decreases during training, eventually converging to approximately 0.5. Figure 12 (e) indicates that the model has learned effectively. The final classification confusion matrix is ​​as follows: Figure 12 As shown in f, the highlighted section on its diagonal indicates that the model can correctly distinguish most categories.

[0073] Furthermore, it should be understood that after reading the foregoing content of this application, those skilled in the art can make various alterations or modifications to this application, and these equivalent forms also fall within the scope defined by the appended claims.

Claims

1. A D-A type narrow-band-gap conjugated polymer, characterized by, The DA-type narrow bandgap conjugated polymer is synthesized from monomers M1 and M2. The structural formula of the monomer M1 is as follows: ； The structural formula of monomer M2 is as follows: ， Wherein, R1 is a C4-C20 straight-chain or branched alkyl, alkenyl or alkynyl group, wherein the alkyl group is optionally substituted with a halogen, alkoxy or cyano group; R2 is a C4-C30 straight-chain or branched alkyl, cycloalkyl, aryl or aralkyl group; M is SnBu3 or SnMe3; Y is a halogen or sulfonate ester. 2.The D-A type narrow-band-gap conjugated polymer according to claim 1, characterized in that, The structural formula of the DA-type narrow bandgap conjugated polymer is as follows: ， Where n is an integer, and the range is n≧2.

3. The method of producing a D-A type narrow-band-gap conjugated polymer according to claim 1 or 2, characterized by, Includes the following steps: The monomers M1 and M2 were subjected to the Stille coupling reaction in the presence of a palladium catalyst and ligands to obtain the DA-type narrow bandgap conjugated polymer.

4. The production method according to claim 3, characterized by, The Stille coupling reaction is carried out in an organic solvent at a temperature of 110-130°C for 24-72 hours.

5. The production method according to claim 3 or 4, characterized by, The monomer M1 was prepared using the following steps: will be described in detail below. and The etherification reaction is carried out in the presence of a first catalyst in a first organic solvent to obtain ; will be described in detail below. with halogenation in a second organic solvent to obtain ; In the presence of a second catalyst, a Yamamoto coupling reaction is carried out in a third organic solvent to obtain ; and ; Will MX3 undergoes a lithiation-tinning reaction with a third catalyst in a fourth organic solvent to obtain the monomer M1. The third catalyst is a lithium catalyst, and X1 and X3 are each independently selected from one of the halogens. Each of the above steps is carried out in an inert gas.

6. The production method according to claim 3 or 4, characterized by, The monomer M2 was prepared using the following steps: In the presence of a fourth catalyst, a Ullmann-type amination reaction is carried out in a fifth organic solvent to obtain and In the presence of a fourth catalyst, a Ullmann-type amination reaction is carried out in a fifth organic solvent to obtain ; Will and An acylation cyclization reaction was carried out in a sixth organic solvent to obtain... ; Will and The Knoevenagel condensation reaction was carried out in the seventh organic solvent to obtain ; Each of the above steps is carried out in an inert gas. Will and The substitution reaction was carried out in an eighth organic solvent to give the monomer M2. In this context, X4 represents a halogen, and Y represents a halogen or a sulfonate.

7. A photoelectric memristor device, characterized in that, include: First electrode; Second electrode; as well as An active layer is disposed between the first electrode and the second electrode; The active layer comprises a polymer layer formed from the DA-type narrow bandgap conjugated polymer as described in claim 1 or 2.

8. The photoelectric memristor device according to claim 7, characterized in that, The active layer further includes a metal oxide layer, and the polymer layer and the metal oxide layer form a heterojunction structure.

9. An application of the photoelectric memristor device as described in any one of claims 5-8 in constructing a neuromorphic computing system, characterized in that, The neuromorphic computing system is a reservoir computing system, which is used for image recognition.

10. The application according to claim 9, characterized in that, In the reservoir computing system, the weight update of the readout layer is achieved by a fully optical modulation method, wherein the weight enhancement is achieved by applying infrared light pulses, and the weight suppression is achieved by applying ultraviolet light pulses.

Images