The following information details the basic operation of a confocal microsocpe. You can leap to any section by clicking on the headings below

Introduction
Limitations
Removing Out of Focus Light
Image Generation
Detection
Pinhole Effects
Slices/Z Planes
Imaging 3D Objects
Representing Data

Introduction

Confocal microscopy is a powerful tool for taking optical slices though a fluorescently stained sample. Samples are still stained with fluorescent probes (antibodies, fusion proteins or dyes) and are visualised using excitation and emission filter systems.  It works on the same principles as standard widefield fluorescence microscopy with three major differences. The first difference is the excitation source, in widefield fluorescent microscopy a mercury or xenon arc lamp is used, for confocal a laser of the specific excitation wavelength required is used. The second difference is that the fluorescent image is collected through a small hole (called the pinhole) that allows only in focus light to pass, resulting in a thin slice of the whole sample being captured. Thirdly the detector is not a camera but a photon multiplier tube. The sample (or the objective) can be moved up and/or down to move the focal plan, resulting in different image slices being generated.

Limitations 

Confocal microscopy is a powerful tool for imaging; however there are limitations to what it can achieve. While confocal microscopy can image thick specimens, it is not going to be able to image a whole leg. The limit of single photon confocal imaging in normal tissue samples is about 150µm. This depth can be drastically reduced, or increased, by many factors; mounting media, fluorophores used, objective used, laser power, sample type, coverslip thickness to name a few.
Resolution is improved on a confocal microscope compared to a widefield microscope, for example a 1.4NA objective imaging GFP will resolve ~180nm in confocal but only ~220nm in widefield. This is only a small increase and means objects can still be 180nm apart and look like one object. As a result “co-localisation” analysis must always be taken with a grain of salt and not be the only basis for a conclusion.
Confocal microscopy with standard PMT detectors is not highly sensitive. The general rule of thumb is, if you can’t see it down the ocular the confocal will struggle to see it as well.
Fluorophores can easily be bleached under confocal scanning, zooming in or increasing the laser power to high levels to detect a signal will greatly increase the chance of bleaching a fluorophore. Also remember that a confocal captures a series of slices, so while the fluorescence maybe fine at the start of acquisition, by the end of the Z stack it may all be gone.

Removing Out of Focus Light 

When a fluorescent sample is excited, excitation and emission does not occur purely at the focal point. Fluorophores above and below the focal point will be excited and emit fluorescence, though not as efficiently as those in the focal plane. When viewed with a widefield microscope these samples look blurred or fuzzy, confocal microscopy removes this by only allowing detection of in focus light.

Basic System Layout – The confocal microscope is essentially the same as a widefield fluorescence microscope. Excitation light is passed to the sample through an excitation filter, is reflected of a dichroic mirror, passed back through the mirror, an emission filter and detected. Excitation is provided by a laser of a specified wavelength and detection is carried out by a photon multiplier tube (PMT). A small hole (called the pinhole) is placed in the emission light path to remove out of focus light.

Excitation – Excitation is provided by a laser of a specified wavelength. The laser is passed through an exitation filter (not shown) and reflected of a dichroic mirror. The light is then passed through the objective to the sample. The excitation light is focused by the objective, but not perfectly. Some out of focus light is applied to the sample as well; this can result in excitation of fluorophores outside of the focal plane.

In Focus Light – Light that is from the plane of focus passes through the dichroic mirror, an emission filter (not shown) and finally the pinhole. Only light that is from the focal plane can pass through the pinhole and thus through to the detector. The resulting image shows only in focus light and is therefore a thin slice of the whole sample. Increasing the size of the pinhole results more light from outside the focal plane being detected, giving a thicker slice. If the pinhole is fully opened the resulting image is essentially the same as that captured with a widefield microscope.

Out of Focus Light – Light that is from outside the plane of focus (either above or below) cannot focus on the pinhole and is therefore not passed through to the detector

Image Generation 

To generate an image the excitation laser is scanned across the sample in a Raster pattern (Raster scanning is the generation of an image one line at a time). The laser sweeps across the sample, is turned off, returned to the starting side of the scan, moved down a set distance and scanned across again. This pattern is repeated until an image of the required size is generated (usually 512 lines).

The laser is scanned by an XY scan unit consisting of a pair of galvo mirrors. The galvo mirrors are moved in a precise fashion, deflecting the beam of another set of mirrors allowing the beam point to be moved either horizontally (X) or vertically (Y). The excitation light is passed through an excitation filter and reflected off a dichroic mirror as described above. Before reaching the objective the beam is passed through the scan unit, this allows the creation of the raster scan pattern to generate the image.

The excitation light emitted from the sample is passed back through the scanning unit, through the dichroic mirror and emission filter. Out of focus light is rejected by the pinhole and the resulting image is detected by the PMT. This form of detection is known as de-scanned detection (meaning the emission light is fed back through the scanning system, or “de-scanned”.
NOTE: Multi-photon confocal systems use non-de-scanned detection. The emission light is detected by PMTs that are placed very close behind the objective to increase sensitivity.

Detection

Light from the sample (photons) is detected by a photon multiplier tube (PMT). A PMT converts collected photons to electrons by a photocathode. The resulting electron passes through a series of gates called dynodes; each dynode carries an electrical charge (between 500 and 1000 volts). Each time an electron contacts a dynode more electrons are produced. The end result of the cascade is many electrons being detected by the anode as a result of one photon entering the PMT.

Increasing the voltage applied to the dynode (called gain, smart gain, HV etc by different manufacturers) results in increased sensitivity of detection, but with an increase in noise as well.
PMTs are not very efficient, approximately 20% of the photons from a sample will be detected. For comparison a high end CCD camera has an efficiency of nearly 70%, a EM-CCD camera can be better than 95% efficient.

Pinhole Effects

The image below demonstrates the effect that opening the pinhole has on the thickness of the optical section captured.
The image is of a piece of fluorescent paper, the individual fibres of which can be easily seen.
The top left image shows only a few fibres that are all in the same plane. As the pinhole size is increased more fibres can be seen above and below the original fibres.

Increasing the size of the pinhole also allows more light to reach the detector, resulting in a brighter image. The example below shows an image generated with the pinhole fully open, and then the pinhole is progressively closed without adjusting for the decrease in brightness. The final image (far right) is barely visible due to the drastically reduced amount of light being detected.

Slices/Z Planes

Confocal microscopy data is usually collected as a series of slices commonly called a Z stack. The distance between each slice is determined when the capture is set up. The example to the right shows a series of slices taken through a sample of fluorescent paper. The total thickness of the collected data is 68µm captured in 2.5µm slices.
This data can then be represented in a range of ways.

Imaging 3D Objects

The vast majority of samples imaged on a confocal microscope will be thick. To be able to effectively image a 3D object the user needs to be aware of what is being captured and what they want to be able to demonstrate with the data.

The example below shows three different slices taken through a sphere, no single slice provides enough information to determine the original objects shape. By looking at all three slices it can be determined that the object is fatter in the middle than on the ends, but what occurs between each of the slices is unknown (there maybe lobes on the object that are not detected).

To accurately represent the sample enough slices need to be taken. If the sample is thin and the fluorescence stable enough then slices that are very close together (or overlapping) can be captured.

But if the sample is thick, fluorescence isn’t stable or time is important then less slices will have to be captured. As a result fine details and some structures maybe lost.

Representing Data

Since confocal data is collected as a series of slices it can be used to generate a range of images. While a single slice, or projection of several slices, may show the result required, sometimes other data representation can be more informative.

All the slices can be combined to generate a maximum intensity projection. This image draws the brightest pixel from each layer in the final image. The result is somewhat like a widefield microscope image with sharper detail

An orthogonal projection can be generated that shows depth information at the selected point. The two thin panels show a view from the right side of the object (YZ dimension) and from the bottom end of the object (XZ dimension). The horizontal and vertical lines in the main image show where the slice has been taken to generate the YZ and XZ views

The slices can be combined to generate a 3D (or 4D if it is a timelapse series) model. This model can be rotated to any angle allowing the user to see parts of the data that may otherwise be missed. Once 3D rendered a data set can be measured for volume, distance etc.