Photon Upmanship: Why Multiphoton Imaging Is More than a Gimmick

Winfried Denk and Karel Svoboda than first power of the light intensity.) The possibility of absorbing more than one photon during a single quan-Bell Laboratories Lucent Technologies tum event had been predicted more than 60 years ago (Goeppert-Mayer, 1931), but experimental confirmation Murray Hill, New Jersey 07974 was delayed until after the invention of the laser (Kaiser and Garrett, 1961) because of the extremely high light intensity necessary to generate appreciable 2-photon Introduction excitation. The following comparison illustrates this point. In bright sunlight, a molecule of rhodamine B, an The unique niche that light microscopy occupies in biol-excellent 1-or 2-photon absorber, absorbs a photon ogy is based on the ability to perform observations on through a 1-photon process about once a second, a living tissue at relatively high spatial resolution. This photon pair by 2-photon absorption every 10 million resolution is limited by the wavelength of light and does years; no 3-photon absorption is expected throughout not rival that of electron microscopy, which is, however, the entire age of the universe. a fundamentally nonvital form of observation. Other vital During the decades following its first demonstration, microscopies, such as MRI, can neither resolve subcel-2-photon absorption was mainly used for the spectro-lular structures nor provide the exquisite molecular se-scopic study of excited molecular states (Friedrich, lectivity that allows the detection of even single mole-1982). This makes sense in spite of the large experimen-cules in a background of billions of others. tal effort required for 2-photon spectroscopy (Birge, Three-dimensional light microscopy inside living tis-1986) because different quantum–mechanical selection sues has been hampered by the degradation of resolu-rules make 2-photon spectra genuinely complementary tion and contrast caused by light scattering, which is to and not just wavelength-scaled versions of their due to refractive index inhomogeneities present to a 1-photon counterparts. This is highly relevant for 2-pho-varying degree in every tissue. Deeper into the tissue, ton microscopy where 2-photon spectra are needed for image degradation becomes progressively more severe choosing appropriate excitation wavelengths. Fortu-and high resolution imaging eventually becomes impos-nately, reliable data are now available for a large number sible. A major step toward overcoming this problem of biologically relevant fluorophores (Xu and Webb, was the invention of confocal microscopy (Minski, 1961), 1996; Xu et al., 1996). Odd (e.g., third) order excitation which uses the resolving power of the objective lens spectra are expected to be more similar to the linear twice: first, the …