Wide-field multiphoton microscopy

The main characteristic of the technique is the illumination of a wide area on the sample with a pulsed laser beam. In nonlinear optics the amount of nonlinear photons (N) generated by a pulsed beam per (illuminating) area per second is proportional to[6][7]

${\displaystyle N\varpropto {\frac {E^{2}}{\tau A}}f}$,

where E is the energy of the beam in Joules, τ is the duration of the pulse in seconds, A is the illuminating area in square meters, and f is the repetition rate of the pulsed beam in Hertz. Increasing the illumination area thus reduces the amount of generated nonlinear photons unless the energy is increased. Optical damage depends on the energy density, i.e. peak intensity per area I_p=E/(τA). Therefore, both the area and energy can be easily increased without the risk of optical damage if the peak intensity per area is kept low, and yet a gain in the amount of generated nonlinear photons can be obtained because of the quadratic dependence. For example, increasing both the area and energy 1000 fold, leaves the peak intensity unchanged but increases the generated nonlinear photons by 1000 fold. This 1000 extra photons are indeed generated over a larger area. In imaging this means that the extra 1000 photons are spread over the image, which at first might not seem an advantage over multiphoton scanning microscopy. The advantage however becomes evident when the size of the image and the scanning time are considered.[3] The amount of nonlinear photons per image frame per second generated by a wide-field multiphoton microscope compared to a scanning multiphoton microscope is given by[3]

${\displaystyle {\frac {N_{\mathrm {wide-field} }}{N_{\mathrm {scanning} }}}=n{\frac {f_{\mathrm {wide-field} }}{f_{\mathrm {scanning} }}}}$,

when assuming that the same peak intensity is used in both systems. Here n is the number of scanning points such that ${\textstyle A_{\mathrm {wide-field} }=nA_{\mathrm {scanning} }}$.

There is the technical difficulty of achieving a large illumination area without destroying the imaging optics. One approach is the so-called spatiotemporal focusing[4][5] in which the pulsed beam is spatially dispersed by a diffraction grating forming a 'rainbow' beam that is subsequently focused by an objective lens.[5] The effect of focusing the 'rainbow' beam while imaging the diffraction grating forces the different wavelengths to overlap at the focal plane of the objective lens. The different wavelengths then only interfere at the overlapping volume, if no further spatial or temporal dispersion is introduced, so that the intense pulsed illumination is retrieved and capable of yielding cross-sectioned images. The axial resolution is typically 2-3 µm[4][5] even with structured illumination techniques.[10][11] The spatial dispersion generated by the diffraction grating ensures that the energy in the laser is spread over a wider area in the objective lens, hence reducing the possibility of damaging the lens itself.

In contrast to what was initially thought, temporal focusing is remarkably robust to scattering.[12] Its ability to penetrate through turbid media with minimal speckle was used in optogenetics, enabling photo-excitation of arbitrary light patterns through tissue.[12] Temporal focusing was later combined with single-pixel detection to overcome the effect of scattering on fluorescent photons.[13] This technique, called TRAFIX, enabled wide-field imaging through biological tissue at great depths with higher signal-to-background ratio and lower photobleaching than standard point-scanning two-photon microscopy.[13]

Another simpler method consists of two beams that are loosely focused and overlapped onto an area (~100 µm) on the sample.[2][3] With this method it is possible to have access to all the elements of the ${\displaystyle \chi ^{(2)}}$ tensor thanks to the capability of being able to change the polarisation of each beam independently.