Abstract
Stimulated emission depletion (STED) microscopy achieves optical super-resolution by suppressing fluorescence at the periphery of the focal spot. While an inverse square root relationship between STED power and resolution is widely accepted, it has historically relied on empirical coefficients. This paper provides a rigorous analytical derivation for the resolution of high-aperture STED systems using $0–2\pi$ vortex phase modulation. By applying Richards-Wolf vectorial diffraction theory and polynomial approximations, the authors derive a precise formula that allows researchers to estimate resolution directly from physical system parameters and dye saturation properties, removing the need for arbitrary fitting constants.
Introduction
In STED microscopy, a "doughnut-shaped" depletion beam is superimposed on an excitation beam to switch off fluorescence everywhere except at the very center of the focal spot. This effectively shrinks the point spread function (PSF) below Abbe’s diffraction limit. However, accurately modeling this process in high-numerical aperture (high-NA) systems—which are standard in super-resolution—requires vectorial optics rather than simple scalar approximations. Previous models either relied on complex numerical simulations that lack intuitive clarity or scalar formulas that fail to account for polarization effects in tightly focused light.
Theoretical Framework and Methodology
The authors bridge the gap between complex simulations and empirical laws by utilizing the Richards-Wolf vectorial diffraction theory. This approach accounts for the polarization states and high-angle rays inherent in high-NA objectives.
Key steps in the derivation include:
- Vectorial Modeling: Modeling circularly polarized incident light combined with a $0–2\pi$ vortex phase mask.
- Polynomial Approximation: To obtain a closed-form analytical expression, the authors approximated the intensity distributions near the focal center using second-order polynomials of the radius ($\rho$).
- Coefficient Derivation: The resulting coefficients in the resolution formula are defined using incomplete Beta functions, which characterize how the high-angle rays contribute to the focal spot's shape.
The Role of Chirality
A critical insight from the vectorial analysis is that the quality of the depletion "doughnut" is highly dependent on the relationship between the light's circular polarization and the vortex phase plate. A perfect zero-intensity center—essential for maintaining signal at the center of the PSF—is only achieved when the chirality of the circular polarization matches the direction of the vortex phase modulation. Discrepancies lead to distorted or filled-in centers, as illustrated in the depletion field profiles.
Analytical Resolution Formula
The paper mathematically validates the empirical square root law while replacing arbitrary constants with defined physical coefficients ($\alpha$ and $\beta$) derived from the system's aperture angle. The derived analytical expression for the Full Width at Half Maximum (FWHM) is:
$$FWHM = \frac{\lambda_{ex}}{\alpha NA \sqrt{1 + \beta I_{dep}/I_s}}$$
For a standard setup ($\lambda_{ex}=635nm, NA=1.4$), the formula yields specific constants. In a Continuous Wave (CW-STED) system, the resolution can be estimated as:
$$FWHM \approx \frac{\lambda_{ex}}{2.15 NA \sqrt{1 + 0.168 I_{dep}/I_s}}$$
Validation against numerical simulations (using the STED3D code and PSFlab) showed that this analytical model maintains less than 5% error when the depletion intensity is high ($I > 10I_s$) and the resolution is significantly below the diffraction limit ($< \lambda/3$).
Conclusion
This work provides the first rigorous analytical description of STED resolution for the widely used $0–2\pi$ vortex phase modulation in high-NA systems. By clarifying the relationship between saturation power ($P_s$) and saturation intensity ($I_s$), and providing a formula based on physical constants, the authors enable experimentalists to predict and optimize system performance without relying on empirical curve fitting. While the model is currently optimized for the vicinity of the focus ($r < 80nm$), it establishes a fundamental theoretical baseline for future developments in super-resolution modeling.
Summary updated