We experimentally demonstrate long-wave infrared-visible sum-frequency generation microscopy for imaging polaritonic resonances of infrared (IR) nanophotonic structures. This nonlinear-optical approach provides direct access to the resonant field enhancement of the polaritonic near fields, while the spatial resolution is limited by the wavelength of the visible sum-frequency signal. As a proof-of-concept, we here study periodic arrays of subdiffractional nanostructures made of 4H-silicon carbide supporting localized surface phonon polaritons. By spatially scanning tightly focused incident beams, we observe excellent sensitivity of the sum-frequency signal to the resonant polaritonic field enhancement, with a much improved spatial resolution determined by visible laser focal size. However, we report that the tight focusing can also induce 1 arXiv:1905.12499v1 [physics.optics] 29 May 2019 sample damage, ultimately limiting the achievable resolution with the scanning probe method. As a perspective approach towards overcoming this limitation, we discuss the concept of using wide-field sum-frequency generation microscopy as a universal experimental tool that would offer long-wave IR super-resolution microscopy with spatial resolution far below the IR diffraction limit.Keywords microscopy, infrared, sum-frequency generation, nanophotonics, surface phonon polariton, polar crystal, semiconductor Nanophotonics relies on the subdiffractional localization of light which is typically achieved using large-momentum surface polariton modes. 1,2 This has been demonstrated for a large variety of structures, using both plasmon polaritons in metallic nanoantennas 3-5 as well as phonon polaritons in subdiffractional polar dielectric nanostructures, 6-13 emerging in the visible (VIS) and long-wave infrared (LWIR) spectral region, respectively. Experimental verification of the resulting light localization, however, is inherently challenging since the spatial resolution in optical microscopes is dictated by the diffraction limit. Specifically in the field of LWIR nanophotonics, this paradigm has been very successfully resolved by using scattering-type scanning near-field optical microscopy (s-SNOM) 14,15 and photothermal induced resonance (PTIR) microscopy, 16,17 where the spatial resolution is limited by the nanoscopic size of the metallic probe tip rather than the imaging wavelength. 9,13,18-28 Further approaches implementing hyperlens concepts may provide additional imaging methods for overcoming the challenge for the diffraction limit as well. [29][30][31] In biology and chemistry, however, the traditional way to achieve image resolution below the diffraction limit has been to harvest nonlinear-optical effects, such as stimulated emission depletion. 32 Infrared (IR) super resolution imaging, on the other hand, has used nonlinear techniques such as IR-VIS sum-frequency generation 33-37 (SFG) or coherent anti-stokes Raman scattering 38,39 microscopy, providing surface-specific and bulk vibrational contrast,