首页|ChemSpecNet: A Neural Network for chemical analysis of Sum Frequency Generation Spectroscopic imaging
ChemSpecNet: A Neural Network for chemical analysis of Sum Frequency Generation Spectroscopic imaging
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NSTL
Elsevier
? 2021 Elsevier B.V.The SFG spectroscopic imaging has been demonstrated as a valuable tool for extracting spatial distributions of the chemical properties in surface chemistry. However, an inherently low signal-to-noise ratio demands pixel binning for extracting reliable chemical information through spectral curve fitting. Consequently, the imaging resolution is reduced. This issue can be circumvented by utilizing Neural Networks (N.N.s). This article demonstrates the utility of the N.N.s by deploying them for chemical identification in the SFG imaging of Self Assembled Monolayers (SAMs) on gold. SAMs are organic molecules that have spontaneously assembled onto a substrate in a single-molecule-thick layer. This new approach solves the chemical identification of the pixel spectra through classification. The method shows high accuracy even with extremely noisy spectra. The resulting chemical labels for the pixels are utilized to reliably generate spatial distributions of the SAMs without pixel binning, thus, enabling full-resolution and fast SFG imaging.
ImagingMicrocontact printingNeural networkSelf assembled monolayersSum frequency generation microscopySurface
NSTL
Elsevier
