2, CEA-Leti, Grenoble, , France
Today, a wide range of AFM-based electrical characterization methods is routinely applied in the study of electronics materials and devices. Methods are available to measure conductivity, charge, surface potential, carrier density, piezo-electric and other electrical properties with nm scale resolution. Typically, these modes are operated in ‘imaging’ or ‘spectroscopic’ mode. In imaging mode, a fixed set of operating conditions (DC & AC sample bias, AC frequency, etc.) are used while the tip is raster-scanning the surface, resulting in high resolution images of height & electrical properties. In spectroscopic mode, the user selects a few points where one of the operating conditions is varied, while the tip is kept stationary. In many studies, both ‘imaging’ and ‘spectroscopic’ experiments complement each other. In this work, the ‘imaging’ and ‘spectroscopic’ modes are combined into a single mode providing an electrical spectrum in every pixel of the image. The tip is moved from pixel to pixel in a fast force volume mapping (FFV) manner, providing a force-distance spectrum in each pixel. During each force-distance cycle, the probe is held on the surface for a pre-defined time at a fixed force or Z position. During this ‘hold segment’, an electrical spectrum is collected by varying one of the electrical operating conditions. This results in a multi-dimensional datacube whereby electrical & mechanical spectra are present for each pixel. Using a force-mapping scan method also overcomes the limitations of contact mode scanning inherent to the conventional implementation of many of the electrical modes: Providing longer tip lifetime and capability to measure soft or fragile samples. Optimization of the force mapping movements, and the electrical measurement hardware & method, allowed us to maintain a relatively high imaging speed (typ. 20-100 ms/pixel). Analysis of the datacubes provides correlation of electrical & mechanical data, and images of a wider range of electrical properties. For example, when collecting an I-V spectrum in every pixel, one can extract a high-resolution map of current barrier properties. This approach is illustrated for Tunneling AFM by ramping the DC sample bias resulting in I-V spectra (FFV-TUNA), for Scanning Capacitance Microscopy by ramping the DC sample bias resulting in dC/dV-V spectra (FFV-SCM), for Scanning Microwave Impedance Microscopy by ramping the DC sample bias resulting in C-V and dC/dV-V spectra (FFV-sMIM), and Piezoforce Microscopy by ramping the DC sample bias (switching loops) or AC frequency (contact resonance spectra) (FFV-PFM), but is also applicable to other electrical modes such as SSRM and electrochemical modes such as SECM. Examples from a variety of materials including semiconductors, ferro-electrics and nanotubes illustrate the capability to reveal sample properties which are not accessible or easily missed in conventional methods where maps at only one or a few discrete settings are acquired.