What is a Beam Splitter?

A beam splitter or power splitter is an optical device that can split an incident light beam e.g. a laser beam into two or sometimes more beams, which may or may not have the same optical power. There are different types of beam splitters; the most important are plate and cube beam splitters as shown in the figure below. Beam splitters are required for various interferometers, autocorrelators, photo cameras, projectors, and laser systems. The wide range of applications implies widely varying requirements, which can be fulfilled with different types of splitters.

Plate Beam Splitters Based on Dielectric Mirrors

Any partially reflecting mirror can be used for splitting light beams as shown in the above figure. In laser technology, dielectric mirrors are often used for such purposes, and they are called plate beam splitters. A dielectric mirror is an optical mirror made of thin layers of dielectric coating layers deposited on an optical substrate. The angle of incidence may be 45° leading to a 90° deflection of one of the output beams, as is often convenient. However, one can design such beam splitters for other deflection angles; they usually work only for a limited range of angles. A wide range of power splitting ratios can be achieved via different designs of the dielectric coating.

The transmitted beam always experiences a spatial shift, the magnitude of which depends on the thickness and the refractive index of the substrate. This is a problem for some applications. Refractive Index is a value calculated from the ratio of the speed of light in a vacuum to that in a second medium of greater density. Generally, the reflectance of a dichroic mirror depends substantially on the polarization state of the beam. Such a device can even be optimized to function as a thin-film polarizer, where in some wavelength range a beam with a certain polarization can be nearly totally reflected, while a beam with different polarization is largely transmitted. On the other hand, it is also possible to optimize for a minimized polarization dependence to obtain a non-polarizing beam splitter within a limited wavelength range.

Dielectric beam splitters usually have a strongly wavelength-dependent reflectance. This can be used for dichroic beam splitters e.g. dichroic mirrors, which can separate spectral components of a beam. The separation may occur based on the difference in wavelength or polarization. A beam splitter as shown above will always lead to a transverse offset of the transmitted beam, which is proportional to the thickness of the used substrate. There are pellicle beam splitters with a very thin substrate, minimizing that beam offset. However, that parasitic reflection from the back side which occurs even if that side is anti-reflection coated may lead to disturbing interferences, and therefore it is often better to use some larger thickness so that the two reflections are spatially well separated.

Beam Splitter Cubes

Many beam splitters have the form of a cube, where the beam separation occurs at an interface within the cube as shown in the above figure. Such a cube is made of two triangular glass prisms which are glued together with some transparent resin or cement. The thickness of that layer can be used to adjust the power splitting ratio for a given wavelength. One may also use some dielectric multilayer coating or a thin metal coating on one or both of the prisms to modify the optical properties, e.g. in terms of operation bandwidth or polarizing properties. As the interface between the prisms is typically very thin, there is only a minimal transverse offset of the transmitted beam. For some applications, this is advantageous, possibly a reason not to use a partially transparent mirror at 45° as shown in the figure. Beam splitter cubes can be used for simple light beams, and also for beams carrying images, e.g. in various types of cameras and projectors.

Cube beam splitters cannot tolerate high optical powers as plate beam splitters, although optically contacted cubes can also exhibit substantial power handling capabilities. Concerning durability and handling, cube beam splitters are often preferred over plates.

Non-polarizing Beam Splitter Cubes

Non-polarizing usually does not imply that such a cube is polarization-preserving. Non-polarizing beam splitter cubes can be made by refining the design, normally via a multilayer coating between the prisms. The substantial angle of incidence will naturally introduce a substantial polarization dependence, but certain design principles can be used to minimize such effects at least within some limited optical bandwidth. For example, if an input beam is polarized at 45° against the axis, it can generally not be expected that the output beam is still linearly polarized since the two polarization components will in general have different phase delays, apart from somewhat different amplitudes.

Polarizing Beam Splitter Cubes

Instead of glass, crystalline media can be used, which can have two different refractive indices. This allows the construction of various types of polarizing beam splitter cubes polarizers such as Wollaston prisms and Nomarski prisms, where the two output beams emerge from the same face, and the angle between these beams is typically between 15° and 45°. Other types are the Glan–Thompson prism, and the Nicol prism, the latter having a rhombohedral form.

Beam Splitters with Geometric Splitting

It is also possible to split beams geometrically, e.g. by inserting a highly reflecting mirror only partially into a light beam, so that some part of the light can pass. One may also use other means, such as the pattern of reflecting stripes or dots on a glass surface. A common design with dots is the Polka dot plate beamsplitter. The advantage of such splitters over dichroic beam splitters is the small wavelength dependence of the splitting ratio. The resulting modification of the intensity profile can be tolerated in some applications (but generally not for imaging).

Beam Splitters with Multiple Outputs

While most beam splitters have only two output ports, there are also beam splitters with multiple outputs. They are fabricated using multiple cascaded beam splitters. Some devices produce some number of output beams of quite similar optical powers with a certain spatial pattern e.g. all in one row, four at the edges of a square, etc.

Fiber-optic Beam Splitters

Various types of fiber couplers can be used as fiber-optic beam splitters. Such a device can be made by fusion-combining fibers and may have two or more output ports. As for bulk devices, the splitting ratio may or may not strongly depend on the wavelength and polarization of the input. Fiber-optic splitters are required for fiber-optic interferometers, as used e.g. for optical coherence tomography. A fiber-optic beam splitter with a single input port and two output ports is shown above. Splitters with many outputs are required for the distribution of data from a single source to many subscribers in a fiber-optic network, e.g. for cable TV.

Other types of beam splitters are: 

• Metal-coated mirrors e.g. half-silvered mirrors, where the metallic coating is made thin enough to obtain partial reflectance
• Pellicles, which are thin membranes, sometimes used in cameras
• Micro-optic beam splitters, often used for generating multiple output beams
• Waveguide beam splitters, used in photonic integrated circuits

Important Properties

Apart from the characteristics concerning the basic function of a beam splitter – the splitting ratio – other properties of beam splitters can be important in applications:

• Some beam splitters are polarizing, while others are non-polarizing. There are also devices designed for use with only one polarization direction – for example, with a laser beam as the input, which is in most cases linearly polarized.
• While some devices work only in a narrow wavelength region e.g. around a common laser line, others are designed for broadband operation, e.g. working throughout the whole visible wavelength region. Similarly, beam splitters may operate properly only with a finite range of incidence angles.
• The optical losses vary significantly between different types of devices. For example, beam splitters with metallic coatings exhibit relatively high losses, whereas devices with dichroic coatings may have negligible losses: the total output power nearly equals the input power. The losses may also be related to the damage threshold, which can be important, particularly for use with Q-switched lasers.
• The spatial configuration can be important for applications. Some require the output ports to be at 0° and 90° relative to the input beam, while others require two parallel outputs or some other configuration.
• For bulk-optical devices, a large open aperture is sometimes needed.

Splitter Applications

Two major examples of the usefulness of beam splitters are:

Emission Image Splitter

Emission image splitters allow for a single camera to image at multiple wavelengths by splitting the camera sensor into sections and projecting emitted light onto each part of the sensor. The optics within an emission image splitter can be seen in figure below.

The above figure shows the internal workings of an emission image splitter. Light travels from left to right in this diagram as indicated by the big arrow. Emitted light from a sample enters the splitter from the microscope and is split into two channels based on wavelength. These channels are manipulated by mirrors to the same camera sensor but offset so that each channel occupies one-half of the camera sensor. The sensor is split vertically into two halves, and two images are observed from one sample.

These splitters act as an interface between the microscope and the camera, emitted light from the sample passes from the microscope to the splitter, and are split based on wavelength before being projected onto sections of the camera sensor. Splitters can split images two, three, or even four times based on wavelengths, allowing researchers to image multiple fluorophores simultaneously rather than having to switch channels manually or electronically. Examples of emission image splitters can be seen in the figure above.

These splitters all attach to standard C-mount ports on microscopes and offer standard output ports for cameras. This makes these splitters a seamless interface, allowing for multiple images on a single camera.

Emission image splitters have a wide variety of uses, as they can split an image across a camera sensor based solely on wavelength, polarization, or amplitude. The main advantage is the simultaneous imaging at multiple wavelengths. If researchers want to image two different fluorophores, typical imaging systems involve manually or electronically cycling through filters to image at the desired wavelength. With a splitter, both wavelengths are imaged simultaneously, suitable for long-term experimentation, fast dynamic events, and any imaging setup that involves multiple fluorescent probes. Brightfield and fluorescence can be imaged simultaneously if desired, and splitting based on polarization allows for flexible experimental setups. Advanced microscopy techniques such as voltage/calcium imaging, Förster Resonance Energy Transfer (FRET), spinning disk confocal, and Total Internal Reflection Fluorescence (TIRF) could all benefit from an emission image splitter.

While an emission image splitter can greatly enhance any imaging system by allowing for simultaneous imaging at 2/3/4 different wavelengths, the main disadvantage is that each channel occupies space on the camera sensor. When using a two-way splitter, the camera sensor is cut in half to resolve two images simultaneously. The two images have half the resolution and field of view due to only having access to half a sensor each. With large camera sensors, this isn’t such an issue, but if a four-way splitter is used on a smaller camera sensor, images may not be resolved properly due to only having a small portion of the sensor to interact with.

Multiple Camera Adapter

While an emission image splitter allows for multiple images on a single camera, the multiple-camera adapter does the opposite: allows multiple cameras to image the same sample. As seen in figure below, a single splitter sends half the light (reflected) from the microscope to one camera, and the other half (transmitted) to a second camera, split based on wavelength, polarization, or amplitude. This setup allows for multiple cameras on one microscope port.

Different multiple-camera adapters allow for increasing numbers of cameras on the same port, with up to four cameras imaging the same sample. These devices can be seen in the figure below.

Multiple camera adapters are shown above where the top splitter is the TwinCam, using a single mirror splitter to allow up to two cameras on one microscope port. The bottom splitter is the MultiCam, using two mirror splitters to allow up to four cameras on one microscope port. These multiple cameras can simultaneously image the same sample.

These splitters work as an interface between the microscope and the cameras, opening up a single microscope port for use with as many cameras as desired. While using an emission image splitter involves decreasing the camera sensor size for each image to acquire images simultaneously, there is no such trade-off with the multiple camera adapter, both cameras retain their full field of view and resolution but can each image the same sample at a different wavelength. This allows for the full potential of each camera to be utilized. In addition, if imaging at challenging wavelengths (such as ultraviolet (UV) or near-infrared (IR)), one camera can be used to cover these wavelengths, while still imaging at conventional fluorescence wavelengths with another camera.

The main drawback is the need to buy multiple cameras as well as a splitter, making the multiple-camera adapter the more expensive option. Despite this additional cost, the significant increase in resolution and field of view make the multi-camera adapters an attractive option for simultaneous imaging.