Confocal Microscopy in a Nutshell
Tips & Tricks for Confocal Imaging
- Fig. 1: Parameters affecting the images quality in a confocal image. Hela cells were fixed and stained with DAPI, Phalloidin-Alexa 488 (Thermo Fischer Scientific) and Mito Tracker Red (Thermo Fischer Scientific). Images were acquired on a Zeiss LSM710 with a 63 x objective NA 1.4. The acquisition parameters were varied as indicated in the different images.
In the last decades confocal microscopes have turned into an invaluable tool in life sciences. Their ability to block the out of focus light (thereby increasing the contrast) facilitated 3D imaging and helped to solve a plethora of scientific questions which could not have been addressed with wide-field microscopy. However the operation of such an instrument requires expert knowledge which is rarely found in scientific publications. We present a generic workflow which can be considered as a beginners guide to obtain a meaningful confocal images enabling state of the art research.
The use of wide-field fluorescence microscopes gets complicated in the case of thick specimen with a rather high labelling density. In such cases the light from the focal plane is outcompeted by the background light from the out of focus planes. As a result the contrast is reduced and especially small and/or dim objects tend to disappear in the sea of background light.
The use of a confocal microscope allows blocking the unwanted background light by adding a small aperture (a so called pinhole) in front of the detector. Classical image formation is replaced by an intensity point measurement. The image is obtained via a reconstruction of multiple point measurements rastering the specimen.
The contrast enhancement seen in a confocal microscope is a consequence of the pinhole in front of the detector (blocking light from out of focus planes) and the scanning process (only one spot in the xy-plane is illuminated). Additionally it can be found that the Point Spread Function (PSF) of the confocal image considerably shrinks along the z-axis and is even mildly improved in the xy plane. Therefore a confocal microscope is also providing a certain resolution improvement. However in order to make use of the improved imaging capabilities of such a system it is necessary to handle the different acquisition parameters of the system correctly. In the following we will discuss these parameters and mention optimization strategies.
The image quality is mainly determined by the signal to noise ratio (SNR) and the signal to background ratio (SBR).
The latter one depending mainly on sample properties (labelling density, labelling specificity) and the acquisition mechanism. The SNR depends on the inherent photon noise (shot noise) and the detector noise. If the detector noise is neglected or at least considered to small compared to the shot noise the quality of an image is scaling with the number of photons which are detected per pixel. Any measure increasing the number of detected photons will increase the SNR (in fact the SNR is proportional to the square root of detected photons).
A slower scan speed increases the so called pixel dwell time and automatically the SNR as the number of detected photons is linearly depending on the time photons are collected. For an image with a dimension of 512 x 512 pixel it is necessary to make 262.000 point measurements resulting in an acquisition time of around 1s per frame if the pixel dwell time is set to 4 µs. In practice a compromise has to be found between image quality on one hand and total acquisition time on the other hand. Pixel dwell times between 6 and 12 µs turned out to be reasonable parameters.
Averaging can be seen as an equivalent of a longer pixel dwell times. It has the advantage of reducing the photobleaching probability as the fluorophore is not being excited all the time and has the chance to go back to the ground state. For objects which are moving averaging is however less well suited.
Due to the spot scanning acquisition mechanism very high light intensities are observed when acquiring a confocal image. As a result the tendency of fluorophores to bleach is in general much higher compared to a wide-field instrument. Even the effect of fluorophore saturation can be encountered on a confocal setup. Bleaching needs to be avoided as it is happening not only in the focal plane but also above and below. Especially 3D imaging is only possible if bleaching is reduced to a minimal level. Therefore any sign of bleaching should always result in decreasing the excitation intensity. In general it is advisable to reduce the excitation intensity as much as possible and only increase it in case if the optimization using other parameters did not lead to a satisfactory result/image.
The pinhole defines the spatial resolution. It mainly affect the axial resolution but also increases the lateral resolution. It is typically given in airy units (AU) and it scales with the wavelength and the NA of the objective. The smaller the AU the more out of focus light is rejected. The optical slices gets thinner. In order to observe a resolution improvement also in the xy-plane. The pinhole needs to be set below 1 AU. As a consequence the number of photons reaching the detector decreases tremendously. So efficient detectors are beneficial for such an approach. In general 1 AU seems to be are very good compromise and starting point for imaging.
The achievable resolution in a confocal image is defined by the PSF and depends on the used wavelength and the NA of the objective. It shall be chosen by the scientific question as a higher resolution does not automatically provide more information. The sampling frequency is an independent parameter which can be rather freely chosen on a confocal microscope. It scales with the zooming factor and the image size. More pixels and a higher zoom result in a higher sampling frequency and thus smaller pixel. The recommended sampling frequency is defined by the Nyquist-Shannon theorem stating that the smallest resolvable structure (in this case defined by the optical resolution limit) must be sampled (at least) twice in order to restore all necessary information.
Typically a photomultiplier (PMT) is used as a detection device. Incoming photons are converted to primary electrons, accelerated in an electrical field and producing secondary electrons which are the read-out. The gain of the PMT is a critical parameter throughout the image optimization. Values between 600 V and 800 V turned out to be a good compromise. Higher values lead to noise images with a reduced dynamic range. Lower values requires rather high excitation intensities and lead to higher photobleaching rates.
In order to make use of the full dynamic range provided by the detector it is advisable to assign the brightest pixel in the image values near the maximal grey value (e.g. 255 for an 8-bit image). This can be typically done by fine tuning the gain value. However care has to be taken to avoid saturation the image (larger areas with the highest pixel value). So called Hi-Low lookup tables are a very efficient tool to avoid saturation.
Main goal when acquiring a confocal image is to balance the different parameters of the instrument being able to answer the scientific question correctly. This is typically a process where a compromise between the SNR and the bleaching rate has to defined and set. Higher SNR favors photobleaching, which in particular needs to be avoided for 3 D imaging.
Authors & Affiliation
Thierry Laroche1, José Artacho1, Arne Seitz1
1EPFL, BioImaging & Optics platform (BIOP), Faculty of Life Sciences (SV), Lausanne, Switzerland
Dr. Arne Seitz
BioImaging & Optics platform (BIOP)
Faculty of Life sciences (SV)