Benefits

TIRFM provides several benefits over standard widefield and confocal fluorescence microscopy such as:

• The background is substantially decreased so structures can be seen clearly
• There is virtually no out-of-focus fluorescence collected which decrease blurring effects
• Cells are exposed to a significantly smaller amount of light which limits phototoxicity to cells

Objective-based vs prism-based

Key differences between objective-based (cis) and prism-based (trans) TIRFM are that prism based TIRFM requires usage of a prism/solution interface to generate the evanescent field, while objective-based TIRFM does not require a prism and utilizes a cover slip/solution interface to generate the evanescent field. Typically objective-based TIRFM are more popularly used, however have lowered imaging quality due to stray light noise within the evanescent wave.

Prism-based

• Less extraneous scattering
• Much cheaper (hundreds instead of thousands of dollars)
• Best for low-mid range magnification and water immersion objectives
• Easiest with free collimated laser sources
• Larger range of incidence angles
  • Desirable to achieve smallest evanescent field depth

(Cis-)total internal reflection fluorescence microscope (TIRFM) diagram

1. Specimen
2. Evanescent wave range
3. Cover slip
4. Immersion oil
5. Objective
6. Emission beam (signal)
7. Excitation beam

Fundamental physics

TIRFM is predicated on the optical phenomena of total internal reflection, in which waves arriving at a medium interface do not transmit into medium 2 but are completely reflected back into medium 1. Total internal reflection requires medium 2 to have a lower refractive index than medium 1, and for the waves must be incident at sufficiently oblique angles on the interface. An observed phenomena accompanying total internal reflection is the evanescent wave, which spatially extends away perpendicularly from the interface into medium 2, and decays exponentially, as a factor of wavelength, refractive index, and incident angle. It is the evanescent wave which is used to achieve increased excitation of the fluorophores close to the surface of the sample, and diminished excitation of superfluous fluorophores within solution.

For practical purposes, in objective based TIRF, medium 1 is typically a high refractive index glass coverslip, and medium 2 is the sample in solution with a lower refractive index. There may be immersion oil between the lens and the glass coverslip to prevent significant refraction through air.

Evanescent wave

The critical angle for excitatory light incidence can be derived from Snell's law:

${\displaystyle \theta _{c}=\sin ^{-1}\left({\frac {n_{1}}{n_{2}}}\right)}$

For ${\displaystyle n_{1}}$ the refractive index of sample, ${\displaystyle n_{2}}$ the refractive index of the cover slip.

Thus, as the angle of incidence reaches ${\displaystyle \theta _{c}}$, we begin observing effects of total internal reflection and evanescent wave, and as it surpasses ${\displaystyle \theta _{c}}$ these effects are more prevalent.

The intensity of the evanescent wave is given by:

${\displaystyle I(Z)=I_{0}e^{-z/d}}$

With penetration depth ${\displaystyle d}$ given by:

${\displaystyle d={\frac {\lambda _{0}}{4\pi }}\left(n_{2}^{2}\sin ^{2}\theta -n_{1}^{2}\right)^{1/2}}$

Typically, ${\displaystyle d}$ ≤~100 nanometers, which is typically much smaller than the wavelength of light, and much thinner than a slice from confocal microscopes.^[1][12]

For TIRFM imaging the wavelength of the excitation beam ${\displaystyle \lambda _{0}}$ within the sample can be selected for by filtering. Additionally, the range of incident angles ${\displaystyle \theta }$ is determined by the numerical aperture (NA) of the objective, and requires that NA > ${\displaystyle n}$. This parameter can be adjusted by changing the angle the excitation beam enters the objective lens. Finally, the reflective indices (${\displaystyle n}$) of the solution and cover slip can be experimentally found or reported by manufacturers.

Excitation beam

For complex fluoroscope microscopy techniques, lasers are the preferred light source as they are highly uniform, intense, and near-monochromatic. However, it is noted that ARC LAMP light sources and other types of sources may also work. Typically the wavelength of excitation beam is designated by the requirements of the fluorophores within the sample, with most common excitation wavelengths being in the 400–700 nm range for biological samples.

In practice, a lightbox will generate a high intensity multichromatic laser, which will then be filtered to allow the desired wavelengths through to excite the sample. For objective-based TIRFM, the excitation beam and fluoresced emission beam will be captured via the same objective lens. Thus, to split the beams, a dichromatic mirror is used to reflect the incoming excitation beam towards the objective lens, and allow the emission beam to pass through into the detector. Additional filtering may be required to further separate emission and excitation wavelengths.

Emission beam

When excited with specific wavelengths of light, fluorophore dyes will reemit light at longer wavelengths (which contain less energy). In the context of TIRFM, only fluorophores close to the interface will be readily excited by the evanescent field, while those past ~100 nm will be highly attenuated. Light emitted by the fluorophores will be undirected, and thus will pass through the objective lens at varying locations with varying intensities. This signal will then pass through the dichromatic mirror and onward to the detector.

Cover slip and immersion oil

Glass cover slips typically have a reflective index around ${\displaystyle n=1.52}$, while the immersion oil refractive index is a comparable ${\displaystyle n=1.51}$. The medium of air, which has a refractive index of ${\displaystyle n=1.00}$, would cause refraction of the excitation beam between the objective and the coverslip, thus the oil is used to buffer the region and prevent superfluous interface interactions before the beam reaches the interface between coverslip and sample.

Objective lens

The objective lens numerical aperture (NA) specifies the range of range of angles over which the system can accept or emit light.

To achieve the greatest incident angles, it is desirable to pass light at an off-axis angle through the peripheries of the lens.

Sample

The sample should be adsorbed to the surface of the glass cover slide and stained with appropriate fluorophores to resolve the features desired within the sample. This is in protocol with any other fluorescent microscopy technique.

Dichroic (dichromatic) filter

The dichroic filter is an edge filter used at an oblique angle of incidence (typically 45°) to efficiently reflect light in the excitation band and to transmit light in the emission band. The 45° angle of the filter separates the path of the excitation and emission beam. The filter is composed of a complex system of multiple layers of metals, metal salts and dielectrics which have been vacuum-deposited onto thin glass.^[13] This coating is designed to have high reflectivity for shorter wavelengths and high transmission for longer wavelengths.^[14] While the filter transmits the selected excitation light (shorter wavelength) through the objective and onto the plane of the specimen, it also passes emission fluorescence light (longer wavelength) to the barrier filter and reflecting any scattered excitation light back in the direction of the laser source.^[15] This maximizes the amount of exciting radiation passing through the filter and emitted fluorescence beam that is detected by the detector.^[16]

Barrier filter

The barrier filter mainly blocks off undesired wavelengths, especially shorter excitation light wavelengths. It is typically a bandpass filter that passes only the wavelengths emitted by the fluorophore and blocks all undesired light outside this band. More modern microscopes enable the barrier filter to be changed according to the wavelength of the fluorophore's specific emission.^[13]

Image artifact and noise

Most fluorescence imaging techniques exhibit background noise due to illuminating and reconstructing large slices (in the z-direction) of the samples. Since TIRFM uses an evanescent wave to fluoresce a thin slice of the sample, there is inherently less background noise and artifacts. However, there are still other noises and artifacts such as poisson noise, optical aberrations, photobleaching, and other fluorescence molecules.

Poissonian noise are fundamental uncertainties with the measurement of light. This will cause uncertainties during the detection of fluorescence photons.^[18] If N photons are measured in a particular measurement, there is a 63% probability that the true average value is in the range between N +√N and N −√N.^[19] This noise may cause misrepresentation of the object at incorrect pixel locations.

Optical aberrations can arise from diffraction of fluorescence light or microscope and objective misalignment. Diffraction of light on the sample slide can spread the fluorescence signal and result in blurring in the convoluted images. Similarly, if there is a misalignment between the objective lens, filter, and detector, the excitation or emission beam may not be in focus and can cause blurring in the images.^[14]

Photobleaching can occur when the covalent or noncovalent bonds in the fluorophores are destructed by the excitation light and can no longer fluoresce.^[20] The fluorescing substances will always degrade to some extent by the energy of the exciting radiation and will cause the fluorescence to fade and result in a dark blurry image.^[21] Photobleaching is inevitable but can be minimized by avoiding unwanted light exposure and using immersion oils to minimize light scattering.^[13]

Autofluorescence can occur in certain cell structures where the natural compound in the structure would fluoresce after being excited at relatively shorter wavelengths (similar to that of the excitation wavelength).^[18] Induced fluorescence can also occur when certain non-autofluorescent compounds become fluorescent after binding to certain chemicals (such as formaldehyde).^[13] These fluorescence can result in artifacts or background noise in the image. Noise from other fluorescence compounds can be effectively eliminated by using filters to capture the desired fluorescence wavelength, or by making sure the autofluorescence compounds are not present in the sample.

Current and future work

Modern fluorescence techniques attempt to incorporate methods to eliminate some blurring and noises. Optical aberrations are generally deterministic (it is constant throughout the image process and across different samples).^[18] Deterministic blurring can be eliminated by deconvoluting the signal and subtracting the known artifact. The deconvolution technique is simply using an inverse fourier transform to obtain the original fluorescence signal and remove the artifact.^[19]

Nevertheless, deconvolution has only been shown to work if there is a strong fluorescence signal or when the noise is clearly identified. In addition, deconvolution performs poorly because it does not include statistical information and can not reduce non-deterministic noise such as poissonian noise. To obtain better image resolution and quality, researchers have used statistical techniques to model the probability where photons may be distributed on the detector.^[18][22] This technique, called the maximum likelihood method, is being further improved by algorithms to improve its performance speed.^[22]