Quantitative coherent Raman scattering microscopies: new tools for the material and life sciences

Abstract

Unlike optical microscopies that are based on fluorescence detection, Raman‐based micro‐spectroscopies provide vibrational signatures that themselves represent quantitative measures of the sample's molecular composition and structures, which for example can be successfully exploited as an intrinsic vibrational contrast of endogenous biomolecular species for label‐free tumour diagnostic imaging [1]. In particular, by exploiting the coherent driving and detection of Raman modes in coherent anti‐Stokes Raman scattering (CARS) and stimulated Raman scattering (SRS), coherent Raman scattering (CRS) microscopy allows the point‐by‐point chemical mapping of molecular compounds, which is often difficult to attain by conventional fluorescence and incoherent vibrational microscopy techniques. Here, we will review on two CRS modalities that provide quantitative molecular information [2]: (i) high‐speed stimulated Raman loss (SRL) imaging at video‐rates and (ii) hyperspectral CARS imaging that provides access to the full wealth of chemical and physical structure information of an a priori unknown molecular sample. We will discuss their underlying principles, their state‐of‐the‐art experimental realizations, and demonstrate exemplifying applications for the label‐free and noninvasive 3D visualization of chemical composition as well as of molecular structure properties of (bio)molecular components in heterogeneous and complex materials, ranging from polymers to living cells. Particular emphasis will be given to the combination of coherent Raman scattering spectroscopy with optical microscopy, which has emerged as a highly sensitive and chemically selective tool for the extraction of quantitative molecular structure information from purely imaginary hyperspectral data cubes of the sample's complex third‐order nonlinear susceptibility, χ (3) (ν,x,y,z), as obtained by fast hyperspectral CARS imaging in conjunction with spectral phase retrieval algorithms. We will introduce a novel concept based on the wavelet prism decomposition and the maximum entropy method (MEM) for the fast and robust reconstruction of the pure vibrational response of the molecular sample inside a sub‐femtoliter probe volume in the presence of experimental artefacts, which may obstruct the accurate phase retrieval from the experimental normalized CARS pixel spectra [3]. Furthermore, we will present exemplifying applications (see Fig. 1) for the quantitative 3D mapping of chemical composition in living cells, the intracellular chemical structure analysis of biologically relevant lipids, and the physical 3D structure analysis in polymers.