Enhancing optical microscopy illumination to enable quantitative imaging

Abstract

There has been an increasing push to derive quantitative measurements using optical microscopes. While several aspects of microscopy have been identified to enhance quantitative imaging, non-uniform angular illumination asymmetry (ANILAS) across the field-of-view is an important factor that has been largely overlooked. Non-uniform ANILAS results in loss of imaging precision and can lead to, for example, less reliability in medical diagnoses. We use ANILAS maps to demonstrate that objective lens design, illumination wavelength and location of the aperture diaphragm are significant factors that contribute to illumination aberrations. To extract the best performance from an optical microscope, the combination of all these factors must be optimized for each objective lens. This requires the capability to optimally align the aperture diaphragm in the axial direction. Such optimization enhances the quantitative imaging accuracy of optical microscopes and can benefit applications in important areas such as biotechnology, optical metrology, and nanotechnology.

Figure 1

The effect of AD axial location on illumination distortions measured at the sample plane (a). Simplified schematic of a reflection (epi) type optical microscope. The AD (or stop) can be moved in the X, Y, and Z directions, as indicated. Measured ANILAS maps for objective type 1 (50×, NA = 0.55), with the AD at axial locations of (b) 2000 μm, (c) 500 μm and (d) −1500 μm. Which can be seen in the supplementary information Video S1. Animation showing changes in the ANILAS map as a function of the AD axial location. Schematic depiction of variations in the angular illumination at the sample plane is also presented at the bottom left.

Figure 2

The relationship among objective type, wavelength, and the axial location of AD on the quality of illumination as characterized by ANILAS map curvature (AMC). The minimum AMC location represents the best quality illumination even though it deviates from the manufacturer-provided AD axial location. Plots are shown of AMC as a function of AD axial location for various objective types at illumination wavelengths of (a) 405 nm, (b) 520 nm, and (c) 633 nm. The objective type, magnification, and NA are shown in the legend in that order. Data points (filled circles) have been fitted with cubic spline curves. The dashed vertical line represents the original axial location of the AD. Error bars represent one standard deviation.

Figure 3

Illumination distortions as represented by the ANILAS maps with the AD located at the original axial location (Z = 0 um) for the illumination wavelength 633 nm. The objective type, magnification, NA, and AMC (x10⁻⁶) values are provided in the headers in that order. The FOV for (a) and (b) is approximately 100 μm, whereas for (c) it is approximately 50 μm.

Figure 4

This figure shows how to select the optimum conditions to obtain illumination of the least distortion. The plot shows the AMC axial location for minimum distortion as a function of illumination wavelength for each of the three objectives. The horizontal blue dashed line shows the axial location of the original AD. Filled circles represent the measured data points. Quadratic curves were fitted to the data points. The solid arrows pointing down from the blue dashed line represent the approximate wavelengths at which the minimum AMC coincides with the original aperture axial location. The objective number, magnification and NA are shown in the legend in that order.

Figure 5

This figure demonstrates the detrimental effect of ANILAS caused by illumination variations upon the TSOM images. ANILAS maps and TSOM images are shown for objective type 4 (50x, NA = 0.95) with the AD located at the original axial position (Z = 0). (a) A typical optical image of the isolated Si line grating on Si substrate used to capture the TSOM images. Pitch is 20 μm, nominal line width is 1 μm, nominal line height is 100 nm. The TSOM images were captured at the locations indicated by the red dots. (b,c) and (d) are the measured ANILAS maps for the illumination wavelengths 405 nm, 520 nm, and 633 nm, respectively. The extracted TSOM image locations with respect to the ANILAS maps are also indicated by the red dots in (b), (c) and (d). (b1, b2), (c1, c2) and (d1, d2) are the captured TSOM images at the illumination wavelengths of 405 nm, 520 nm and 633 nm, respectively. The red dashed lines indicate the tilt of the TSOM image axis.

Figure 6

This figure demonstrates the detrimental effect of illumination distortions upon the intensity profiles. Shown here are variations in the intensity profiles within the FOV as well as the dependence of these variations upon the degree of illumination quality and upon the choice of focus position. The intensity profiles are for isolated Si line gratings on a Si background with nominally similar dimensions and symmetric cross sectional profiles (nominal line width is 1000 nm, nominal line height is 100 nm, pitch is 20 μm). An objective lens of type 6 (100x, NA = 0.85) was used with λ  = 633 nm for panels (a) and (b) while a type 4 (50x, NA = 0.95) was used with λ  = 405 nm for panels (c) and (d). The position of the AD was at the original AD axial location (Z = 0). Thus panels (a) and (b) were obtained with a strong ANILAS presence (Fig. 3c), whereas (c) and (d) were obtained using a good-quality illumination (Fig. 5b). Profiles in panels (a) and (c) were extracted by placing targets close to the position of best focus. Profiles in panels (b) and (d) were extracted by moving the targets axially 1.8 μm closer towards the objective than they were for panels (a) and (c). Optical intensities were normalized to simplify the comparison.