Quantitative Fluorescence Microscopy

Quantitative Fluorescence Microscopy Frank R. Boddeke promotor: Prof. dr. I.T. Young toegevoegd promotor: Prof. dr. ir. L.J. van Vliet -------------------------------------------------------------------------------- Summary This thesis discusses several issues that are central to instrumentation in quantitative wide-field fluorescence microscopy. A general introduction and a overview of this thesis is given in Chapter 1. Chapter 2 discusses autofocusing in microscopy. First a model of the image formation and image acquisition is described. This model shows the important relation between the magnification of the microscope, the sampling of the camera and the optical transfer function (OTF) of the microscope optics. The sampling of the camera is given by the spacing of the photo-sensitive elements in the camera. The OTF is determined by the numerical aperture (NA) of the objective lens and the wavelength of the light used. This model leads to two simple focus criteria to be used in different sampling situations: the {1,-1} filter is optimal in case of sampling at the Nyquist frequency, the{1, 0, -1} filter is optimal in case of sampling at half the Nyquist rate. Where reducing the sampling to half the Nyquist rate (e.g. by on-chip binning of the CCD pixels) results in aliassing when considering image formation, focusing -with the correct criterion- benefits from this due to noise (more light per pixel) and speed (less pixels) improvements. Using the latter preferred focus criterion a three-phase focus algorithm has been designed. The first two phases (course and fine focusing) step through the z-axis with certain step sizes. The third phase (refine focusing) samples the focus function in a small z-range around focus. The final in-focus z-position is then derived from a quadratic fit through these samples. The size of the quadratic region is derived from the NA of the objective lens. With this approach we found that the accuracy of the focusing procedure is smaller than the reproducibility of the z-axis (25 nm) for contrast of at least 1600 photo-electrons per pixel for a camera with a noise level of 11.3 photo-electrons. Chapter 3 discusses several calibration techniques for the automated z-axis of a microscope. The z-axis motor step size is determined with use of a tilted slide. Focus functions are used to map changes in z-position (in motor steps) to measurable changes in x-position. If all parameters are known (the tilt of the slide and the sampling density in the x-direction), the relation between a shift in the x and in the z-axis, allows calculation of the z-axis motor step size. The backlash and the stability of the z-axis are also measured using focus functions. The shift between two focus functions (focus values as function of the z-position) -one derived by acquiring images while the stage is moving upwards and one while moving downwards- is the backlash of the z-axis. Monitoring the focus value over time, without moving the stage, and comparing it to a previously acquired focus function, gives insight into the stability of the z-axis. Chapter 4 discusses a wide range of topics concerning image sensors as they are used in quantitative microscopy. An overview of different cameras is given. The chapter focuses however on (grey-scale) charge-coupled devices (CCD) and (second generation) intensified CCD (I-CCD) cameras. The physics of image acquisition is described to allow insight in important physical limitations (limited signal-to-noise ratios due to the quantum nature of light and limited resolution due to the wave description of light). After that the properties of CCD cameras are examined and methods for characterizing CCD cameras are derived. These properties include the noise sources, linearity of response and spatial frequency response. These properties are also examined for image intensifiers. Image intensifiers have some additional properties which have no equivalent in CCD cameras. Such as the spatial and temporal low pass filtering of image intensifiers and its effect on the signal-to-noise ratio. These properties are also examined and a theoretical model is given. Chapter 5 describes the development of a fluorescence lifetime imaging microscope (FLIM). The goal of the research described in this chapter is the development of a prototype commercial product. Special attention has been paid to the cost of the components, the robustness of the system and the ease-of-use. The developed system applies the Fourier method and homodyne detection to detect fluorescence lifetimes in the range of 1 to 100 ns in a standard wide-field fluorescence microscope. The FLIM modulatable light source (constructed of a laser-diode connected to an optical fiber) simply replaces the regular arc-lamp and the FLIM modulatable image sensor (constructed of a modulatable image intensifier and a video camera) replaces the camera on the camera port of a regular fluorescence microscope. A specially designed computer controlled modulation signal generator synthesizes the RF modulation signals (1 to 100 MHz) for the light source and image sensor. Image-processing and control software allow automatic acquisition of fluorescence lifetime images. The prototype system has been built and several tests have been performed to evaluate the quality of the system. Further Information