Epi-fluorescence microscope
One objective both illuminates the sample and collects its fluorescence; a dichroic mirror and two filters keep the excitation out of the camera.
Open in the canvas →Background
A fluorescent molecule absorbs light of one colour and, a few nanoseconds later, emits light of a longer wavelength. The emission is weak — typically many orders of magnitude weaker than the light that excites it — and it leaves in all directions. A fluorescence microscope is therefore mostly an exercise in separating a faint colour from a bright one[1].
The epi arrangement does the first part of that with geometry. The excitation is sent to the sample through the objective, so most of it carries on through the sample and away from the detector; the objective then collects only the fluorescence emitted back towards it, plus whatever excitation the sample scatters or reflects. The rest is done with three coatings, usually mounted together in one cube[1,2]:
an excitation filter that passes only the band the dye absorbs; a dichroic mirror that reflects that band towards the objective and transmits the longer-wavelength emission coming back; and an emission filter in front of the detector that passes the dye's emission band and blocks, by five or six orders of magnitude, the excitation that still got through the dichroic.
The objective is used in the infinity configuration: the light from a point in its focal plane leaves it collimated, crosses the dichroic and the emission filter as a parallel beam — which is why flat plates can be put there without disturbing the image — and a tube lens focuses it on the camera.
The filter set is chosen for the dye. Enhanced GFP, used here, absorbs most strongly near 488 nm and emits with a peak near 507 nm[3], so its excitation and emission bands are only about 20 nm apart and the dichroic edge has to fall between them.
What this setup demonstrates
A 488 nm laser passes a 488/10 excitation filter and is reflected downwards by a long-pass dichroic with its edge at 495 nm. The objective focuses it into a GFP sample on a piezo stage, which scans the sample sideways under the fixed focus.
The sample emits GFP's band in all directions. The part that falls inside the objective's acceptance cone is collected and collimated, passes the dichroic, is trimmed by a 525/50 emission filter, and is focused by a 100 mm tube lens onto the camera. The camera's reading is about a thousandth of the laser power, and its spectrum lies entirely between 500 and 550 nm.
Things to try. Select the camera and look at the wavelength it reports: no 488 nm. Then move the emission filter's centre to 620 nm — clear of GFP's band — and the camera goes dark, because there was never any excitation leaking through for it to pass instead. Move the dichroic edge below 488 nm and the laser goes straight through it and never reaches the sample. Lower the objective's NA and the collected signal falls with the width of the cone.
The fluorescence is qualitative. The sample converts a set fraction of the excitation that reaches it into the dye's emission band and radiates it isotropically in the plane of the drawing; there is no quantum yield, extinction coefficient, concentration, saturation or bleaching, and the thousandth that reaches the camera is a 2D collection fraction, not the solid angle of a real objective.
The filters and the dichroic are ideal: full transmission inside the band, none outside, edges that do not move with angle. So the excitation rejection here is perfect, where a real filter set is specified by how many orders of magnitude it achieves, and the sample neither reflects nor scatters the laser back. Autofluorescence and background are absent.
There is no image. The 6 mm laser beam underfills the objective's 10 mm pupil and is brought to a focus; nothing models diffraction, the point-spread function, resolution, or the camera as a 2D sensor. The stage motion is a display animation of the sample under the focus, and a uniform sample gives the same reading everywhere.
Related components
References
- J. W. Lichtman, J.-A. Conchello, “Fluorescence microscopy,” Nature Methods 2, 910–919 (2005)
- Nikon MicroscopyU — Introduction to Fluorescence Microscopy
- R. Y. Tsien, “The green fluorescent protein,” Annual Review of Biochemistry 67, 509–544 (1998)