Multiphoton Microscopy

Multiphoton microscopy

Multiphoton microscopy (MPM) is an imaging technique that allows intact biological samples to be visualized at high resolution. MPM is based on nonlinear optical processes such as two-photon and three-photon excited fluorescence (2PEF, 3PEF) and second and third harmonic generation (SHG, THG) (see figure 1 below). Most commonly, 2PEF and 3PEF are used to image fluorescent dyes or endogenous molecules. SHG is used to image non-centrosymmetric strucutres such as collagen fibers and THG is used to visualize the refractive index differences such as the interface of lipid droplets.

Multiphoton microscopy has several advantages over microscopy. Due to the nonlinear nature, multiphoton signal generation is essentially confined to the focal volume, giving high resolution without the need for spatial filtering as in confocal microscopy. Additionally, photo bleaching of fluorescent molecules outside the focus is virtually eliminated. Another important advantage of multiphoton absorption over single-photon absorption is that near infrared light is used instead of visible. Longer wavelength light experiences less scattering and absorption in biological tissue, allowing for deeper penetration.

Figure 1. Corregistered multimodal MPM images of fresh human breast cancer tissue and energy level diagrams for different optical processes. Standard fluorescence imaging relies on 1-photon excited fluorescence where a single higher energy photon is absorbed, followed by emission of a lower energy photon. Non-linear 2-photon (3-photon) excited fluorescence is based on the simultaneous absorption of two (three) lower energy photons. Second (third) harmonic generation is another nonlinear optical process where two (three) photons are converted to a single photon without losing any energy to the environment. Coherent anti-Stokes Raman scattering imaging is a third-order nonlinear optical process in which the vibration of chemical bonds is coherently amplified. More description of CARS can be found below.

Multiphoton microscopes are typically based on Titanium sapphire lasers. These types of lasers produce ultrashort pulses of light that are necessary to excite non-linear optical processes. Figure 2 shows the typical schematics of a multiphoton microscope. The Titanium –sapphire laser beam (red) is focused into the sample by a high numerical aperture objective lens. The MPM signal (green) is collected by the same objective and directed to a detector by a dichroic mirror. The beam is scanned across the sample to create the MPM image. 

In addition to the typical use of Titanium sapphire lasers, there have been tremendous efforts in the development of novel lights sources for multiphoton imaging in order to improve the imaging performance and to reduce the cost. By using fiber sources with longer wavelength, lower repetition rate, broad bandwidth and dispersion-corrected pulses (Figure 3), we have significantly improved the signal generation efficiency of the multiphoton processes and were able to speed up the label-free imaging of breast cancer in situ or even in vivo while achieving optimized molecular contrast as shown in Figure 1.

Figure 2: Schematic of a typical multiphoton system based on Titanium sapphire laser.


Figure 3 Schematic of a fiber-based multi-photon system. Different dichroic mirrors (DM) and optical filters are used in the detection system to collect spectrally resolved multimodal multiphoton signals by photomultipliers (PMT).

Coherent anti-Stokes Raman scattering (CARS)

CARS can stimulate the production of a significantly larger amount of signal than spontaneous Raman microscopy. Like spontaneous Raman, CARS probes vibrational modes in molecules and does not require the introduction of exogenous dyes or markers, which is advantageous in imaging small molecules, such as metabolites, for which labeling may significantly affect their molecular properties.

CARS is a process that involves four photons that interact with the third order nonlinear susceptibilityof the sample, which is a function of the vibrational frequencies . To understand a CARS event, consider two photons: a pump, of energy , and a Stokes, of lower energy. Consider also a molecule with a single resonance, represented by a third order susceptibility . A CARS event can be understood in two steps (Figure 1). Upon the illumination of the molecule with the pump and Stokes photons, the first step is initiated if the conditionis met; that is, if the di erence in energy between the pump and Stokes photons matches the energy of the excited vibrational state of the molecule, so that the molecule is excited. Once this happens, the second step is the result of the interaction of this excited state with a third photon, known as the probe, of energy This photon gains the energy of excitation of the molecule, and an anti-Stokes photon is emitted with an energy  that has a higher frequency than any of the incident photons.

The first systematic study of CARS was performed by Maker and Terhune in 1965. Since then, CARS spectroscopy has become a technique widely used in chemical analysis. Duncan et al. reported the construction of the first CARS microscope in 1982. In this experiment, they used two picosecond dye lasers (in the visible portion of the spectrum) in a non-collinear con guration to meet the phase matching condition of the process, and images of the distribution of D2O where taken with the use of a two-dimensional detector, but the signal was overwhelmed by an overlapping background. Following them, in 1999, Zumbusch et al. reported the first microscope in a tightly focused configuration that collected light in the forward direction, this time with collinear pump and Stokes pulses in the near infrared range. The new selection of frequency for the incident light showed a reduction in the background, and the relaxation of the phase-matching condition (due to tight focusing) allowed better spatial resolution. This report triggered more studies of CARS microscopy in the following years.

The following two lists summarize the main advantages and disadvantages of CARS microscopy for biological applications.


  1. CARS microscopy makes use of a nondestructive, noninvasive endogenous contrast mechanism, which is advantageous in imaging small molecules for which labeling may significantly affect the molecular properties. Furthermore, the endogenous contrast mechanism in CARS is not affected by photobleaching and problems with diffusion of markers to the target zones.
  2. As a nonlinear optical process, a CARS signal is generated in zones with high optical intensities. The intensity of the acquired signal has a quadratic dependence on the pump field and a linear dependence on the Stokes field, a condition that limits the generation of the CARS signal over a reduced region in the focal point when high numeric aperture objectives are used. This allows using CARS microscopy in a confocal configuration to produce 3D vibrational contrast images with high spatial resolution.
  3. Imaging requires fast detection. The effciency of CARS is orders of magnitude higher than that of spontaneous Raman scattering. This enables CARS as a good contrast mechanism for imaging with chemical contrast (spectra can be acquired at ~10m/pixel, which is reasonably fast).
  4. When illuminating a sample to produce CARS photons, other processes occur simultaneously, i.e., fluorescence, coherent Stokes Raman scattering, stimulated Raman gain (SRG), stimulated Raman loss (SRL) and Raman-induced Kerr effect (RIKE). CARS signals belong to a separate range of the spectrum from these other processes, so there is no conflict in identifying signals coming from them.


  1. The presence of a coherent nonresonant background can overwhelm the chemically specific resonant signal. Since the signals mix coherently, suppression of the background is not trivial, and many approaches have been taken to solve this problem, one of which is presented as a topic in this work.
  2. For spontaneous Raman microscopy, the molecular oscillators within the sample are random in phase. For CARS, however, the molecular oscillators are stimulated to vibrate in phase, making it necessary to illuminate the sample within certain phase matching conditions. Therefore, CARS signal is generated in specific directions at which constructive interference of the anti-Stokes field occurs.
  3. The experimental implementation of CARS is, in general, costly and diffcult when compared to linear techniques, fluorescence or SHG microscopy. Conventional CARS needs the use of two synchronized pulses, and they need to overlap in space and time at the sample, which requires very precise alignments and apo-chromatic objective lenses.