Novel method for preparing bionic flexible strain sensor

Abstract

According to the novel method for preparing the bionic flexible strain sensor, the preparation parameters of the graphene are predicted by utilizing machine learning in artificial intelligence, so that the experiment efficiency is improved. The laser-induced graphene (LIG) with excellent performance is prepared by utilizing predicted data; in order to increase the toughness of LIG, the prepared LIG is mixed with silica gel (Ecoflex) to prepare an LIG / Ecoflex film. By studying natural bionics, it is determined that a spider web structure serves as a bionic structure, a novel construction method of the spider web structure is provided, and according to the method, a carbon dioxide laser and a fiber laser (C-FL for abbreviation) are adopted for combined action; the C-FL method can solve the problems that a carbon dioxide laser construction structure is rough and a fine structure is constructed in a small area. The whole sensor is simple in manufacturing process, economical and durable, and can be prepared on a large scale. The device can be used for detecting stretching and bending of human joints, and has certain potential in the directions of medical treatment, motion detection, man-machine interaction and the like.

Description

Technical Field

[0001] Flexible strain sensors Background Art

[0002] Flexible sensors have become a hot topic in recent years. Their convenience, rapid response, high precision, and excellent durability make them promising for widespread application in a wide range of fields. Flexible sensors consist of a surface encapsulation layer, a sensitive layer, and a substrate. The sensitive material is the primary component of the sensor. Currently, a variety of sensitive materials exist, including carbon materials, metal nanoparticles, conductive fabrics, and liquid metals. However, among these materials, graphene, due to its unique properties, is considered the most promising for sensor manufacturing.

[0003] In 2014, Jian Lin discovered that using a CO2 infrared laser to ablate commercial polyimide (PI) films could directly generate patterned three-dimensional porous graphene without generating hazardous waste. This process, known as laser-induced graphene (LIG), is widely used in the fabrication of flexible sensors due to its simple preparation method and controllable graphene structure. LIG has shown great potential in the fields of healthcare and human signal detection. LIG materials have been widely used in sensor manufacturing. For example, Qiong Tian et al. designed a self-healing flexible pressure sensor with a peapod-like structure. This pressure sensor exhibits ultrahigh sensitivity and linearity, with a sensing range of up to 100 kPa and stable performance in cycling tests exceeding 1000 cycles. Chhetry et al. generated MDS-LIG with excellent electromechanical properties by laser ablation on a MoS2-coated PI surface. The sensor, fabricated based on MDS-LIG, exhibits high sensitivity and stability, and can be used to detect subtle skin deformations such as wrist pulse, sound, and finger movement. Jeong et al. generated LIG on a PI surface using laser irradiation and then transplanted it onto a PDMS surface for encapsulation, creating a flexible, highly sensitive strain sensor with excellent response. Previous work has shown that microstructural modifications can improve sensor performance, but the process is complex and difficult to implement. Furthermore, graphene vacancies lead to high sensor resistance, limiting performance. Therefore, to address these issues, we propose modifying the graphene structure at the macroscale rather than at the microstructure. Research has revealed that the structure of a natural spider web is an octagonal network composed of numerous radial threads radiating from a center and spiral threads surrounding the center. This structure has a wide coverage area, is interconnected end to end, and exhibits excellent resonant sensing capabilities. When used in sensor design, this structure induces structural resonance wherever strain occurs, thereby improving sensor sensitivity and stability. However, spider webs are delicate and difficult to fabricate. To achieve a near-replica of the spider web structure, we employed a C-FL method to construct the spider web structure. This method is convenient, can be directly constructed using a laser, and the resulting structure is indistinguishable from the original spider web.

[0004] Machine learning is a discipline that uses artificial intelligence computers as tools, encompassing multiple disciplines such as statistics and probability theory. It uses theoretical knowledge and complex algorithms to simulate real human working methods, thereby improving experimental efficiency. Machine learning has widespread application across various fields, improving the efficiency and accuracy of various tasks by simulating human learning and decision-making processes. The combination of machine learning and sensors has broad application. In addition to traditional human-computer interaction devices such as keyboards, mice, and touchscreens, new human-computer interaction devices based on speech recognition, gesture recognition, eye tracking, and brain-computer interfaces have also been developed. Xie et al. constructed a voice recognition sensor integrated with machine learning, using machine learning to identify sensor data and achieve accurate speech recognition. Machine learning applications in sensors are also reflected in data analysis. Zhang et al. used machine learning to analyze and train known data from a multi-region wireless sensor network. They then predicted missing attack data in the target region, successfully predicting security event trends in unknown areas. This paper leverages the ability of machine learning to predict and optimize unknown data using known data, successfully predicting LIG fabrication parameters. Summary of the Invention

[0005] This sensor uses a CO2 laser to directly produce LIG as the sensitive material. To improve efficiency, we successfully predicted LIG production parameters using machine learning. First, we obtained a certain amount of experimental data through experiments and preprocessed and analyzed the data. Then, we selected an appropriate neural network model based on the data's characteristics and distribution, and trained and optimized the model using historical data. Finally, we validated and evaluated the model using test data, achieving satisfactory prediction results. The predicted results were used to successfully prepare graphene with excellent performance. However, the toughness of LIG prepared directly on the PI substrate is poor. To solve this problem, we proposed to transfer the LIG formed on the PI substrate to the Ecoflex substrate. The three-dimensional porous structure of graphene allows Ecoflex to effectively penetrate into the pores, thereby improving its stability and toughness; thus, the corresponding structure can be constructed on its surface. Through research, we found that the LIG / Ecoflex membrane has a spider web structure. However, the diameter of the carbon dioxide laser spot is 760μm, and the required structure cannot be directly constructed on the LIG / Ecoflex membrane. To solve this problem, we proposed a new method to construct a spider web structure. This method uses a carbon dioxide laser (CO2 Laser) and a fiber laser (Fiber Laser). The team used a C-FL method to construct a spider web structure, successfully creating a small, distinct spider web structure on the LIG / Ecoflex membrane. The C-FL method perfectly creates a spider web structure on the LIG / Ecoflex membrane, resulting in a flexible spider web sensor with good durability, reversibility, stability, fast response time (200ms) and recovery time (260ms) within the bending deformation range. Furthermore, the sensor has the advantages of being small in size but with a clear structure, simple preparation, and low cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 Schematic diagram of the overall structure of the present invention Figure 2 Material characterization diagram prepared using predicted values ​​for this invention Figure 3 This is the performance test diagram of the present invention Figure 4 This is the preparation flow chart of the present invention Figure 5 This is a practical application diagram of the present invention.

Claims

1. A bionic spider web flexible strain sensor prepared using a new method includes a surface encapsulation layer, a sensing layer with a spider web structure, and a flexible substrate. Its characteristics are that the sensing material is composed of a laser-induced polyimide (PI) film, and the spider web structure sensing layer and the substrate can be prepared together.

2. The spider web strain sensor according to claim 1, wherein In the construction of the spider web structure, a carbon dioxide laser is used to scan the polyimide (PI) film to prepare laser-induced graphene. The spider web structure is constructed using a fiber laser with a smaller spot diameter. This method is called the C-FL method. Using this method, the spider web structure can be constructed within an area of ​​10mm*10mm and the constructed structure is clear.

3. The spider web strain sensor according to claim 2, characterized in that The preparation method of laser-induced graphene adopts machine learning, which uses machine learning to predict the preparation parameters of laser-induced graphene.

Images