Single photon avalanche diode (SPAD) array detectors for luminescence based biomedical applications
Single photon avalanche diode (SPAD) array detectors, with their single photon sensitivity and picosecond time resolution, are particularly well suited for detecting the small number of photons produced in a chemiluminescence measurement and for determining fluorescence lifetime, which is typically in the nanosecond range. In this dissertation, the key parameters that SPAD array detectors needs to have in order to be used for luminescence-based biomedical applications were investigated. Two application areas were investigated, one being the detection of pathogens in a compact chemiluminescence-based measurement system and the other being fluorophore differentiation using fluorescence lifetime in a flow cytometry measurement system.
Using experimental measurements on two measurement setups developed for this purpose and suitable Monte Carlo simulations enabled the detection of SARS-CoV-2 RNA and ssDNA at different concentrations. It was shown that the detection limit defined as acceptable by the WHO ( 106 copies/ml) can theoretically be achieved with this setup and a suitable assay of 1.8⋅105 copies/ml. The main factors were also determined to further lower the detection limit. Using the second measurement setup, the next step was to investigate the accuracy and precision of fluorescence lifetime determination and its use for fluorophore differentiation. For this purpose, the relation of the accuracy and precision of the determined fluorescence lifetime to the photon counts and the detection rate were investigated, and the required parameters of the SPAD array detector, such as pixel count and time resolution, were determined. The pile-up effect, which limits measurement duration and decreases with higher pixel number, can be reduced with as few as
30 pixels, so that the fluorescence lifetime can already be determined in microsecond measurement durations.
In cases such as pathogen differentiation and characterization, differentiation of a large number of fluorophores in a mixture is necessary. Here, differentiation by fractional contributions using fluorescence lifetimes provides an additional multiplexing factor, when the limit of spectrally differentiable fluorophores is reached. It was shown that the number of differentiable fluorophores is mainly limited by the difference between the fluorescence lifetimes present, which must be at least 0.5 ns for two fluorophores. By using artificial neural networks, the accuracy and precision of the determined fractional contributions could be further increased compared to the conventionally used LS-Fit method. To verify this method for use in flow cytometry, single fluorescent microparticles were differentiated despite high background fluorescence fraction of about 50 % at a fluorescence lifetimes difference of 0.7 ns. From these measurement results it can be concluded that it is theoretically possible to differentiate about 5000 cells or particles per second with the SPAD array detector used here. The requirements for the measurement setup and the SPAD array detector were successfully determined for both applications investigated in this dissertation, enabling the development of customized SPAD array detectors.
Using experimental measurements on two measurement setups developed for this purpose and suitable Monte Carlo simulations enabled the detection of SARS-CoV-2 RNA and ssDNA at different concentrations. It was shown that the detection limit defined as acceptable by the WHO ( 106 copies/ml) can theoretically be achieved with this setup and a suitable assay of 1.8⋅105 copies/ml. The main factors were also determined to further lower the detection limit. Using the second measurement setup, the next step was to investigate the accuracy and precision of fluorescence lifetime determination and its use for fluorophore differentiation. For this purpose, the relation of the accuracy and precision of the determined fluorescence lifetime to the photon counts and the detection rate were investigated, and the required parameters of the SPAD array detector, such as pixel count and time resolution, were determined. The pile-up effect, which limits measurement duration and decreases with higher pixel number, can be reduced with as few as
30 pixels, so that the fluorescence lifetime can already be determined in microsecond measurement durations.
In cases such as pathogen differentiation and characterization, differentiation of a large number of fluorophores in a mixture is necessary. Here, differentiation by fractional contributions using fluorescence lifetimes provides an additional multiplexing factor, when the limit of spectrally differentiable fluorophores is reached. It was shown that the number of differentiable fluorophores is mainly limited by the difference between the fluorescence lifetimes present, which must be at least 0.5 ns for two fluorophores. By using artificial neural networks, the accuracy and precision of the determined fractional contributions could be further increased compared to the conventionally used LS-Fit method. To verify this method for use in flow cytometry, single fluorescent microparticles were differentiated despite high background fluorescence fraction of about 50 % at a fluorescence lifetimes difference of 0.7 ns. From these measurement results it can be concluded that it is theoretically possible to differentiate about 5000 cells or particles per second with the SPAD array detector used here. The requirements for the measurement setup and the SPAD array detector were successfully determined for both applications investigated in this dissertation, enabling the development of customized SPAD array detectors.