Transmission and transflection infrared microscopy of biological cells and tissue suffer from significant baseline distortions due to scattering effects, predominantly resonant Mie scattering (RMieS). This scattering can also distort peak shapes and apparent peak positions making interpretation difficult and often unreliable. A correction algorithm, the resonant Mie scattering extended multiplicative signal correction (RMieS-EMSC), has been developed that can be used to remove these distortions. The correction algorithm has two key user defined parameters that influence the accuracy of the correction. The first is the number of iterations used to obtain the best outcome. The second is the choice of the initial reference spectrum required for the fitting procedure. The choice of these parameters influences computational time. This is not a major concern when correcting individual spectra or small data sets of a few hundred spectra but becomes much more significant when correcting spectra from infrared images obtained using large focal plane array detectors which may contain tens of thousands of spectra. In this paper we show that, classification of images from tissue can be achieved easily with a few (<10) iterations but a reliable interpretation of the biochemical differences between classes could require more iterations. Regarding the choice of reference spectrum, it is apparent that the more similar it is to the pure absorption spectrum of the sample, the fewer iterations required to obtain an accurate corrected spectrum. Importantly however, we show that using three different non-ideal reference spectra, the same unique correction solution can be obtained.

Transflection-mode FTIR spectroscopy has become a popular method of measuring spectra from biomedical and other samples due to the relative low cost of substrates compared to transmission windows, and a higher absorbance due to a double pass through the same sample approximately doubling the effective path length. In this publication we state an optical description of samples on multilayer low-e reflective substrates. Using this model we are able to explain in detail the so-called electric-field standing wave effect and rationalise the non-linear change in absorbance with sample thickness. The ramifications of this non-linear change, for imaging and classification systems, where a model is built from tissue sectioned at a particular thickness and compared with tissue of a different thickness are discussed. We show that spectra can be distorted such that classification fails leading to inaccurate tissue segmentation which may have subsequent implications for disease diagnostics applications.

Molecularly Imprinted Polymers (MIPs) are synthetic receptors that are able to selectively bind their target molecule and, for this reason, they are currently employed as recognition elements in sensors. In this work, MIP nanoparticles (nanoMIPs) are produced by solid-phase synthesis for a range of templates with different sizes, including a small molecule (biotin), two peptides (one derived from the epithelial growth factor receptor and vancomycin) and a protein (trypsin). NanoMIPs are then dipcoated on the surface of thermocouples that measure the temperature inside a liquid flow cell. Binding of the template to the MIP layer on the sensitive area of the thermocouple tip blocks the heat-flow from the sensor to the liquid, thereby lowering the overall temperature measured by the thermocouple. This is subsequently correlated to the concentration of the template, enabling measurement of target molecules in the low nanomolar regime. The significant improvement in the limit of detection (a magnitude of three orders compared to previously used MIP microparticles) can be attributed to their high affinity, enhanced conductivity and increased surface-to-volume ratio. It is the first time that these nanosized recognition elements are used in combination with thermal detection, and it is the first report on MIP-based thermal sensors for determining protein levels. The developed thermal sensors have a high selectivity, fast measurement time (<5 min), and data analysis is straightforward, which makes it possible to monitor biomolecules in real-time. The set of biomolecules discussed in this manuscript show that it is possible to cover a range of template molecules regardless of their size, demonstrating the general applicability of the biosensor platform. In addition, with its high commercial potential and biocompatibility of the MIP receptor layer, this is an important step towards sensing assays for diagnostic applications that can be used in vivo.