Optical sectioning: linear versus nonlinear signals
The same objective focusing into the same fluorescent cuvette: one-photon fluorescence glows along the whole beam, two-photon fluorescence only at the focus.
Open in the canvas →Background
Optical sectioning is what lets a microscope image one plane inside a thick specimen without the planes above and below it washing the image out. In this setup only the nonlinear signal is sectioned: the same focused beam excites one-photon fluorescence all along its path, and two-photon fluorescence only at its focus.
Linear signals come from the whole beam
A focused beam carries the same power through every plane along its path; only the area it is spread over changes. A signal that grows in proportion to the intensity — one-photon fluorescence, spontaneous Raman — therefore comes from a thin slice of the specimen equally well wherever the slice sits in the cone, in focus or not. A wide-field fluorescence microscope collects all of it, so out-of-focus light adds a haze to the image. A confocal microscope removes it after the fact: a pinhole in front of the detector, conjugate to the focus, rejects the fluorescence excited above and below the plane in focus[2]. The fluorescence is still generated everywhere along the beam, and so is the photobleaching[2].
Nonlinear signals come from the focus
Two-photon fluorescence needs two photons absorbed together, so it grows with the square of the excitation intensity[1,2]. Through each slice the power is fixed but the intensity falls as the beam widens, so the signal from a slice falls as the beam's area to the power n − 1 for an n-photon process. For a Gaussian focus that is (1 + (z/zR)²)−(n−1) at a distance z from the focus, where zR is the Rayleigh range: a slice a few Rayleigh ranges away already gives a small fraction of what the focal slice gives, and three-photon fluorescence and third harmonic generation (n = 3) fall off faster still. The fluorescence, and the bleaching, are confined to the focal region[1]: the sectioning is intrinsic to the excitation, no pinhole is needed, and every emitted photon can be collected, even scattered ones[2]. This is what Denk, Strickler and Webb showed in the first two-photon laser-scanning microscope[1].
Nonlinear light does not need a focus as such. A wide-field or speckle beam with enough peak power generates it too, but then it comes from everywhere the beam is, and there is no sectioning.
What this setup demonstrates
Both arms are the same: a laser, a laser-line filter, an objective (EFL 10 mm, NA 0.4, working distance 5 mm) and a cuvette of GFP solution with 1 mm glass walls and 6 mm inside, focused 3 mm into it. Only the excitation differs.
- 488 nm, continuous: one-photon fluorescence glows along the whole 6 mm path through the cuvette, including the cone before and after the focus. The workbench emits it from spots every 0.5 mm along the beam, one of them exactly at the focus — the point a confocal pinhole would image.
- 920 nm, 150 fs: two-photon fluorescence comes only from the focus. Select the cuvette: Focus inside gives the focus 3 mm into the active layer with a Rayleigh range of about 2.4 µm.
Things to try. Move the lower cuvette so the focus lies in its glass wall, more than about 75 µm (31.6 Rayleigh ranges) from the inside, and the two-photon signal disappears while the upper one does not change. Replace the objective with a weak lens, and the two-photon glow stretches along the beam as the focus grows longer. Put a diffuser in the beam, or remove the objective, and there is no focus left to section with: the two-photon signal comes from the whole path, like the one-photon one. The specimen's Sample layout also offers a thin layer between two glasses, the usual microscopy sample.
The cuvette is a qualitative layered specimen. Each ray crosses it in a straight line, with no refraction at the glass, and the signal is emitted from a few spots along that line. The focus is an ideal diffraction-limited Gaussian in air, estimated from the convergence of the traced rays: no aberrations, refractive index, scattering or high-NA vector effects. Two-photon light is generated only where the intensity is at least 1/1000 of the focus's, an illustrative limit; a real microscope's sectioning (a few micrometres) is set by the detected signal, not by a cutoff. The one-photon signal grows with the path at its efficiency per millimetre, and the two-photon total is its efficiency, reached when the focal region lies inside the active layer; neither is a cross-section or a peak-intensity calculation. No pinhole or confocal detection is part of this setup.
Related components
References
- W. Denk, J. H. Strickler and W. W. Webb, "Two-photon laser scanning fluorescence microscopy," Science 248, 73–76 (1990)
- P. Luu, S. E. Fraser and F. Schneider, "More than double the fun with two-photon excitation microscopy," Communications Biology 7, 364 (2024)