The scanned focus
A point-scanning microscope focuses a laser to a spot, moves the spot over the sample and records one signal per position. The image of a point object is therefore the spot itself, weighted by what the detector accepts [1][4]. For a uniformly filled circular pupil, the scalar intensity near the focus depends on two dimensionless optical coordinates, one across and one along the axis [2][5]:
Without a pinhole a one-photon signal is proportional to the excitation intensity at the object, so across the focus the point image is h at the excitation wavelength, with FWHM 0.5145 λexc/NA, which the Zeiss monograph rounds to 0.51 (its eq. 3) [1]. Along the axis the excitation focus has FWHM 0.8859 λexc/(n − √(n² − NA²)), the monograph’s 0.88 (eq. 2; 1.77 nλ/NA² for a small NA, eq. 2a), but without a pinhole that is not an axial resolution, as explained below. The page finds each FWHM by bisection on the computed profile and converts optical units to micrometres:
The FWHM is not the same as the two-point resolution: the Rayleigh criterion places two incoherent points at the Airy pattern’s first zero, 0.61 λ/NA apart.
No axial resolution without a pinhole. A one-photon signal is excited in every plane the beam crosses, above and below the focus, and a large detector collects it all: a uniform fluorescent layer returns the same total signal at every depth, so the signal cannot be assigned to a depth [2]. The monograph likewise gives no optical slice thickness for conventional microscopy, where sectioning is not possible [1]. The page therefore reports no axial resolution for one-photon excitation without a pinhole. That is a choice of what to report, not the absence of an axial profile: an isolated point, such as a bead on a dark background, still gives one, the excitation focus’s axial FWHM above; but in a sample that fluoresces above and below the focus every plane adds the same total signal, so that width does not locate the signal in depth [2]. A confocal pinhole selects the focal plane in detection; multiphoton excitation confines the excitation itself to the focus (below).
| Symbol | Quantity | Unit | Meaning |
|---|---|---|---|
| λexc | Excitation wavelength | nm | Vacuum wavelength of the scanning laser. |
| λdet | Detected wavelength | nm | Emission peak, Raman Stokes line, or λexc for scattered light. It matters only with a pinhole. |
| NA | Numerical aperture | — | n sin α of the objective; must be below n. |
| n | Refractive index | — | Immersion medium, taken to equal the sample’s. |
| α | Aperture half-angle | ° | asin(NA/n), the half-angle of the focused cone. |
| PH | Pinhole diameter | AU | Back-projected to the sample, in Airy units of the detected light: 1 AU = 1.22 λdet/NA. |
| N | Photons per event | — | 2 for two-photon fluorescence and SHG, 3 for three-photon fluorescence and THG. |
The confocal pinhole
A pinhole in a plane conjugate to the focus accepts light from a small region around it. The total point image is the product of the illumination and detection point images (Zeiss eq. 1) [1], and the detection point image is the detected-wavelength PSF summed over the back-projected pinhole of radius R [3]:
The two coordinates of hdet scale with λexc/λdet, so a Stokes shift widens the detection side. The page evaluates D numerically for any pinhole between 0 and 10 AU. In the two limits it reproduces the monograph’s closed forms, listed beside the result for comparison [1]:
- Large pinhole (PH > 1 AU), “geometric-optical confocality”: the resolution is that of the excitation spot alone (eqs. 2, 3). The pinhole still rejects light from other planes; the monograph measures that by the optical slice thickness, the axial FWHM of the detection response D (eq. 4).
- Point pinhole (PH < 0.25 AU), “wave-optical confocality”: the total PSF is hexc·hdet, which the monograph approximates with one mean wavelength λ̄ (eqs. 6–8). Below 1 AU it no longer distinguishes the optical slice thickness from the axial resolution: the axial FWHM of the total PSF is both.
Resolution and optical sectioning are different properties [4]. At 1 AU the computed lateral width is within a few per cent of the open-detector value, yet the pinhole already rejects light from other planes; the lateral gain appears only well below 1 AU, where the detected signal falls steeply [1]. The monograph recommends z steps of half the slice thickness.
Spontaneous Raman is a linear, incoherent signal like fluorescence; the detected wavelength is the Stokes line of the chosen shift:
Scattered or reflected light is treated as coming from point-like incoherent scatterers. Extended reflecting structures image coherently: the monograph notes, for example, a factor 0.45 instead of 0.64 in eq. 7 for a mirror [1]. In confocal Raman and reflection work an index mismatch between immersion and sample can make the real depth resolution much worse than these figures [11].
Two- and three-photon excitation, SHG and THG
An N-photon signal grows with the N-th power of the excitation intensity, so the point image is the excitation focus raised to that power [2][6][7]. Detection needs no pinhole: a large, non-descanned detector collects everything.
Raising the profile to a power narrows it: at the same wavelength the two-photon point image is laterally about 1.39 times narrower than the one-photon one (0.3693 against 0.5145 λ/NA), the three-photon image about 1.70 times. A two-photon process is, however, excited at roughly twice the wavelength a one-photon process needs for the same transition, so it is wider than one-photon excitation at λ/N [2][6]. For two- and three-photon fluorescence the page lists both lateral comparisons and adds one-photon confocal imaging of the same dye at λ/N through a pinhole, 1 AU unless changed: the monograph calls about 1 AU the typical setting and, for most fluorescence work, the best compromise between signal and depth discrimination [1]; along the axis, one-photon excitation without a pinhole is drawn as the flat signal of a thin uniform layer. The confocal point image is the narrower one: for the same emission wavelength, two-photon fluorescence does not resolve better than confocal microscopy [17]. Measured confocal resolution, however, often falls short of its theory while nonlinear microscopes come close to theirs, so in practice they show no resolution disadvantage [18]. The reasons to choose two-photon excitation lie elsewhere: excitation confined to the focus bleaches and damages the specimen less outside the imaged plane, the sectioning needs neither a pinhole nor descanning optics, and near-infrared light is scattered and absorbed less, so it reaches deeper into thick tissue [2]. SHG and THG have no one-photon counterpart, so the page compares them with nothing. Horton et al. used three-photon excitation at 1700 nm to image deep in the scattering mouse brain [8].
What multiphoton excitation gains is sectioning without a pinhole: the total signal from a uniform plane falls away from the focus instead of staying constant [2]. That plane response sums hexcN over each plane, so its width is a different quantity from the axial FWHM of the point image, which is what the page reports. Second- and third-harmonic generation (SHG, THG) also follow the square and the cube of the intensity, and are emitted at exactly λ/2 and λ/3; from a point-like object they have the same point image; extended objects add coherent effects (phase matching, the Gouy phase) that are not modelled.
Field of view and galvo scanning
An infinity-corrected objective’s magnification is defined with its maker’s tube lens: 200 mm for Nikon and Leica, 180 mm for Olympus and 165 mm for Zeiss [13]. The objective’s focal length follows from it, and the field of view in the sample is the field number divided by the magnification [14]:
In a laser-scanning microscope a galvanometer mirror tilts the beam, a scan lens turns the tilt into a position in an intermediate image, and the tube lens and the objective image that plane onto the sample [12]. A mirror turned by θmech deflects the beam by twice that angle. An ideal, distortion-free scan lens places the beam at fSL tan θ; an f-theta lens at fSL θ [12]:
Scan and tube lens form a telescope that images the galvo onto the objective’s back aperture and expands the beam by fTL/fSL [12]. The resolution panel assumes a uniformly filled aperture. The Zeiss monograph describes the fill by the truncation ratio T of the 1/e² beam diameter to the pupil diameter: below about 0.6 the focus is Gaussian, above about 1 it is close to the Airy pattern, and beyond 3 the resolution no longer improves while more power is clipped [1].
Raster scanning with two galvos
A frame is scanned line by line. The fast mirror (X) sweeps the focus across one line; the slow mirror (Y) then moves it to the next, and after the last line returns it to the first. Each mirror follows the command voltage from its servo driver, proportional to its rotation (volts per degree, the driver’s scale factor). The page draws both commands for the field set above.
- Unidirectional. The fast command is a sawtooth: a linear ramp, then a quick flyback. Pixels are recorded only while the mirror moves at constant speed; the data taken during the turnaround and the flyback are discarded, so only part of each line period, the fill fraction, makes the image [15].
- Bidirectional (serpentine). The fast command is a triangle and both sweeps record a line, every other one in reverse, so no time is spent on a flyback. A galvo’s inertia makes it lag its command, by about 7 % of the line period in the ScanImage example [15], and the lag grows into image distortion at higher line rates [16]. In a bidirectional image it shifts the forward and the return lines in opposite directions, so the two must be aligned by a delay or phase correction, or vertical edges look jagged.
Here the slow mirror advances one line during each flyback or turnaround and holds still while a line is recorded. The graph shows the commands, not the mirror positions, which lag and round the corners.
Precision and limits
Numerics. Bessel functions use rational approximations accurate to about 10⁻⁸. The pinhole integral uses Gauss–Legendre quadrature (64 radial × 48 angular nodes across the pinhole, 128 nodes over the pupil for the defocused field), and each FWHM is bisected to round-off. An independent Python reference, using integral representations of the Bessel functions and a different quadrature, agrees with the page to 10⁻⁶ relative or better in every case of the validation table (the test tolerance is 10⁻⁵); the open and point-pinhole limits are also checked against their analytic half-maximum points.
Scope. Scalar optics with a uniformly filled, aberration-free circular pupil; the axial coordinate uses n(1 − cos α), exact on the axis, and the defocused field off the axis is the paraxial form. Above about NA/n = 0.7 the vector nature of the focused field (Richards and Wolf [9]) widens the real focus along the polarization, which this page does not include. Emission is taken at a single wavelength; a broad band widens the detection side slightly. The sample is assumed index-matched to the immersion medium, non-scattering and free of aberrations; real foci deep in tissue or through a mismatched index are wider, and the depth scale itself changes [11]. Cole et al. describe how to measure the point image with sub-resolution beads and compare it with such limits [10].
Scanning. Thin, ideal scan and tube lenses in a telescope that images the galvo onto the back aperture; an objective that maps the pupil angle to fobj tan θ; no distortion, field curvature or vignetting. The field number limits where the objective is corrected; the page flags a scanned field beyond it but does not model the loss.
References
- S. Wilhelm, Confocal Laser Scanning Microscopy: Principles, Carl Zeiss Microscopy GmbH, publication 60-1-0030/e (2010); copy hosted by the Technion Biomedical Core Facility.
- P. T. C. So, C. Y. Dong, B. R. Masters and K. M. Berland, “Two-photon excitation fluorescence microscopy,” Annu. Rev. Biomed. Eng. 2, 399–429 (2000).
- T. Wilson and A. R. Carlini, “Size of the detector in confocal imaging systems,” Opt. Lett. 12, 227–229 (1987).
- T. Wilson, “Resolution and optical sectioning in the confocal microscope,” J. Microsc. 244, 113–121 (2011).
- M. Born and E. Wolf, Principles of Optics, 7th ed., Cambridge University Press (1999): Fraunhofer diffraction at a circular aperture and the three-dimensional light distribution near a focus.
- M. Gu and C. J. R. Sheppard, “Comparison of three-dimensional imaging properties between two-photon and single-photon fluorescence microscopy,” J. Microsc. 177, 128–137 (1995).
- C. J. R. Sheppard, “Image formation in three-photon fluorescence microscopy,” Bioimaging 4, 124–128 (1996).
- N. G. Horton, K. Wang, D. Kobat, C. G. Clark, F. W. Wise, C. B. Schaffer and C. Xu, “In vivo three-photon microscopy of subcortical structures within an intact mouse brain,” Nat. Photonics 7, 205–209 (2013).
- B. Richards and E. Wolf, “Electromagnetic diffraction in optical systems, II. Structure of the image field in an aplanatic system,” Proc. R. Soc. Lond. A 253, 358–379 (1959).
- R. W. Cole, T. Jinadasa and C. M. Brown, “Measuring and interpreting point spread functions to determine confocal microscope resolution and ensure quality control,” Nat. Protoc. 6, 1929–1941 (2011).
- N. J. Everall, “Confocal Raman microscopy: why the depth resolution and spatial accuracy can be much worse than you think,” Appl. Spectrosc. 54, 1515–1520 (2000).
- A. Negrean and H. D. Mansvelder, “Optimal lens design and use in laser-scanning microscopy,” Biomed. Opt. Express 5, 1588 (2014).
- K. R. Spring and M. W. Davidson, “Infinity optical systems,” Evident (Olympus) Microscopy Resource Center.
- Nikon, Microscopy glossary: field of view (field number divided by the objective magnification).
- T. A. Pologruto, B. L. Sabatini and K. Svoboda, “ScanImage: flexible software for operating laser scanning microscopes,” BioMed. Eng. OnLine 2, 13 (2003).
- W. Wang, Z. Wu and H. Zeng, “Image distortion and its correction in linear galvanometric mirrors–based laser-scanning microscopy,” J. Biomed. Opt. 20, 056001 (2015).
- C. J. R. Sheppard, “Multiphoton microscopy: a personal historical review, with some future predictions,” J. Biomed. Opt. 25, 014511 (2020).
- G. Cox and C. J. R. Sheppard, “Practical limits of resolution in confocal and non-linear microscopy,” Microsc. Res. Tech. 63, 18–22 (2004).