首页|期刊导航|安徽大学学报(自然科学版)|基于脉冲加热单MOS传感器的二甲苯检测研究

基于脉冲加热单MOS传感器的二甲苯检测研究OA

Research on the detection of xylene based on a single MOS sensor with pulse-heating

中文摘要英文摘要

金属氧化物半导体传感器在苯系物气体检测中普遍存在交叉敏感问题,严重制约其选择性识别能力.针对上述问题,提出一种基于脉冲加热的单传感器策略,利用5 wt%金修饰的锌锡氧化物为敏感材料,采集其在苯、甲苯、二甲苯及两种含二甲苯的二元混合气体(二甲苯+甲苯、二甲苯+苯)中的动态响应信号,提取响应峰值、响应时间、积分响应面积及稳态响应均值4类时域特征参数,结合主成分分析、径向基核函数支持向量机分类算法和随机森林回归算法,建立气体识别模型与二甲苯浓度预测模型,进而开展二甲苯识别与浓度预测研究.结果显示,单组分和二元混合气体的分类准确率均为100%;在二甲苯、二甲苯+甲苯、二甲苯+苯3类回归任务中,测试集决定系数R2分别为0.999 7,0.766 9和0.963 4.研究表明,脉冲加热能够有效激发敏感材料在非平衡态下的差异化动力学响应,显著提升苯系物各组分在时间维度上的可分辨性,为单金属氧化物半导体传感器对二甲苯的选择性识别与浓度预测提供有力支撑.

Metal oxide semiconductor gas sensors commonly suffer from the problem of cross-sensitivity in the detection of benzene series gases,which severely restricts their selective recognition ability.To address the above problem,a single-sensor strategy based on pulse-heating was proposed.A 5 wt%gold-modified zinc-tin oxide was used as the sensitive material,and its dynamic response signals in benzene,toluene,xylene,and two binary mixed gases containing xylene(xylene+toluene,xylene+benzene)were collected.Four types of time-domain characteristic parameters,including response peak value,response time,integral response area,and steady-state response mean value,were extracted.Gas recognition and xylene concentration prediction models were established by combining principal component analysis,radial basis function support vector machine classification algorithm,and random forest regression algorithm,and then the research on xylene recognition and concentration prediction was carried out.The results showed that the classification accuracy for single-component and binary mixed gases reached 100%.In the three regression tasks for xylene,xylene+toluene,and xylene+benzene,the determination coefficients R2 of the test sets were 0.999 7,0.766 9,and 0.963 4,respectively.The research indicated that pulse-heating could effectively stimulate the differential dynamic responses of the sensitive material in the non-equilibrium state,significantly improve the distinguishability of each component of the benzene series in the time dimension,and provide strong support for the selective recognition and concentration prediction of xylene by a single metal oxide semiconductor sensor.

吴庆;徐安麒;李成森;赵康智;唐曦;赵长辉

安徽大学物质科学与信息技术研究院,安徽 合肥 230601安徽大学物质科学与信息技术研究院,安徽 合肥 230601安徽大学物质科学与信息技术研究院,安徽 合肥 230601安徽大学物质科学与信息技术研究院,安徽 合肥 230601安徽大学物质科学与信息技术研究院,安徽 合肥 230601安徽大学物质科学与信息技术研究院,安徽 合肥 230601

信息技术与安全科学

脉冲加热锌锡氧化物二甲苯气体识别浓度预测

pulse-heatingzinc-tin oxidexylenegas recognitionconcentration prediction

《安徽大学学报(自然科学版)》 2026 (5)

57-66,10

安徽省高等学校科学研究重点项目(2023AH050067)

10.3969/j.issn.1000-2162.2026.05.008

评论