Dual Detection of Urea and Glucose in Sweat Using a Portable Microfluidic SERS Sensor with Silver Nano-Tripods and 1D-CNN Model Analysis

• Published in: ACS applied materials & interfaces 2024-12, Vol.16 (48), p.65918-65926
• Main Authors: Xiong, Siyue, Wang, Chengxuan, Zhu, Chushu, Dong, Peitao, Wu, Xuezhong
• Format: Article
• Language: English
• Subjects: Biological and Medical Applications of Materials and Interfaces; Biosensing Techniques - instrumentation; Biosensing Techniques - methods; Glucose - analysis; Humans; Lab-On-A-Chip Devices; Limit of Detection; Metal Nanoparticles - chemistry; Silver - chemistry; Spectrum Analysis, Raman; Sweat - chemistry; Urea - analysis
• ISSN: 1944-8244; 1944-8252; 1944-8252
• DOI: 10.1021/acsami.4c14962

Sweat, a noninvasive metabolic product of normal physiological responses, offers valuable clinical insights into body conditions without causing harm. Key components in sweat, such as urea and glucose, are closely linked to kidney function and blood glucose levels. Portable sweat sensors, equipped with diverse sensing systems, can monitor fluctuations in urea and glucose concentrations, thus providing methods for assessing kidney function and monitoring diabetes. This study presents a flexible, portable microfluidic surface-enhanced Raman scattering (SERS) sensor designed to detect the unique fingerprint of target biomarkers. This flexible, self-adhesive microfluidic chip, constructed from modified polydimethylsiloxane, features silver nanotripods (AgNTs) with densely distributed “hotspots” created via the oblique angle deposition technique. These AgNTs act as active substrates for SERS within the microfluidic platform, enabling direct skin contact to collect, transport, store, and analyze sweat. The chip functions as a quantitative urea sensor with a limit of detection (LOD) of 10–7 M. For enhanced sensitivity for glucose detection, the SERS substrate is modified with 4-mercaptophenylboronic acid, achieving a LOD of 10–7 M. This satisfies the measurement requirements for both urea and glucose in sweat under physiological conditions. Furthermore, the one-dimensional convolutional neural network model significantly enhances the accuracy of biomarker detection, facilitating the quantitative analysis of urea and glucose. This advancement contributes to the development of a controlled, convenient, and dynamic biosensing system for personalized point-of-care testing and supports the creation of intelligent wearable and nondestructive devices.