To shrink device size yet still tightly control performance, new technology often requires increasingly stringent surface specifications. Characterization tools, in turn, must keep pace by providing higher resolution, faster throughput, and more functionality. The atomic force microscope (AFM) is well known as a high-resolution imaging technique, but its characterization power and ease of use have increased significantly over the years.
Principles of AFM Operation
Unlike optical and electron microscopes that “see” a surface via transmitted or reflected radiation, the AFM “feels” the surface using a micromachined cantilever with a small tip (Figure 1). With a typical radius of a few nanometers, this tip allows the AFM to sense surface forces with high sensitivity.
The cantilever's position is controlled in three dimensions by a scanner with piezoelectric actuators. Imaging is accomplished by raster scanning the sample (or equivalently, the cantilever) in the XY plane. To improve scanning accuracy, newer AFMs use closed-loop positioning (i.e., sensored feedback) to compensate for piezoelectric hysteresis and creep. Newer AFMs also feature designs with better mechanical stability that reduce thermal drift and noise. As a result, high-resolution imaging can now be performed without vibration isolation or temperature control systems in many settings.
During scanning, the cantilever is monitored with an optical detection scheme. A laser beam is focused onto the cantilever and reflected into a position-sensitive photodiode. The up-down photodiode voltage captures the cantilever's relative vertical position and can be converted to absolute deflection through established calibration procedures. The left-right photodiode voltage can also be acquired and represents the cantilever's relative lateral or torsional motion.
The AFM controller, or control system, includes active feedback to improve sensitivity. The photodiode deflection signal is input to a feedback loop that controls the cantilever position. An example is contact mode, where the tip scans in contact at constant applied force. In this case, the feedback loop works to maintain constant cantilever deflection, and thus force, using the Z piezo actuator to adjust the height of the cantilever base. Images then represent the height change needed at each position. Nearly all AFM modes of operation utilize feedback control, but the variable controlled and the type of data acquired differ depending on the mode.
Several key features of AFMs are apparent from even this brief description. Nearly any type of material can be examined, and samples usually require little or no preparation. Imaging can be performed in an ambient environment or even in liquid. The small tip size provides spatial resolution far exceeding other stylus methods. However, the AFM's field of view is relatively small, typically a few tens of micrometers. Scan speed has also been slow historically (a few minutes per image), but dramatic increases1 have recently been made (up to several frames per second).
Imaging Topography on the Nanoscale and Beyond
Since their invention in the 1980s, AFMs have used the above concepts to map surface height variations with nanoscale resolution. These height, or topography, images provide valuable information on many types of structure including roughness, defects, amorphous and crystalline phases, and thin-film nucleation and growth.
Topography was originally acquired in contact mode, described above. However, scanning in contact induces lateral forces that can damage delicate samples, cause excessive tip wear, and require slower scanning. To address these issues, an approach called tapping mode was quickly developed. In tapping mode, the cantilever is oscillated at constant frequency near a flexural resonance (typically from tens to hundreds of kilohertz). Oscillation has traditionally been performed with a piezoelectric “shaker,” but photothermal excitation2 and other alternatives have recently been developed.
In tapping mode, variations in the tip-sample interaction force during scanning alter the cantilever's time-averaged oscillation amplitude. The feedback loop works to keep this amplitude constant by adjusting the cantilever's Z position, and the image data represents these adjustments. Because the tip only touches the sample intermittently in tapping mode, lateral forces are greatly reduced and much faster imaging is possible. Tapping mode also yields a second image of cantilever oscillation phase that can provide useful contrast between sample components.
The example topography image in Figure 2 emphasizes that AFM images are 3D surface profiles with quantitative data, not 2D projections subject to interpretation. They thus lend themselves to a wide range of analysis and display options, many of which are pre-programmed in software on newer AFMs. For example, image metrics that correlate with performance or processing variables can be determined quickly and easily.
With an image width of only 10 nm, the topography image in Figure 3 demonstrates the ultra-high resolution possible with today's AFMs. Hardware improvements have enabled spatial resolution sufficient for lattice-scale imaging – resolution similar to, or even better than, the current limit3 of ~50 pm for high-resolution transmission electron microscopy. Since the tip-sample interaction volume determines AFM resolution, its limits are far smaller than those set by optical and electron diffraction.
Besides higher spatial resolution and other technical improvements, today's AFMs are also easier to use than earlier models. New automated routines – for instance, to align the laser or optimize imaging parameters in tapping mode – greatly reduce setup time. Operation is further streamlined by an extensive range of built-in tools for image display and analysis.
Imaging Local Material Properties
The AFM's capabilities go beyond imaging topography, however. The same force-sensing concepts can be used to quantify near-surface physical properties on the nanoscale. For many applications, such measurements provide valuable information that imaging morphology alone cannot.
For example, functional properties such as electrical, magnetic, and electromechanical response impact applications ranging from photovoltaics to nonvolatile memory and data storage. To interrogate functional behavior on the nanoscale, a number of AFM modes4 have been developed based on electrostatic, capacitive, magnetic, and related tip-sample interactions.
AFM modes that probe electrical properties include conductive AFM (CAFM), electrostatic force microscopy (EFM), and Kelvin probe force microscopy (KFPM). The example in Figure 4 shows CAFM evaluation of a photoactive film. The nanoscale information provided by these techniques is often complementary to that obtained by probe station methods, which test a whole device. AFM electrical techniques can also be used to assess uniformity, identify defects, and otherwise assure quality.
Other capabilities for functional characterization are provided by piezoresponse force microscopy (PFM) and magnetic force microscopy (MFM). PFM characterizes static and dynamic electromechanical response of piezoelectric, ferroelectric, and multiferroic materials. In contrast, MFM uses a magnetized tip to assess the magnetic behavior of ferromagnetic and multiferroic materials.
In other applications, mechanical and tribological properties such as modulus, adhesion, and friction are critical for performance and reliability. The AFM's sensitivity to low forces enables mechanical measurements with much higher vertical and lateral resolution than possible otherwise. As Figure 5 shows, today's AFMs provide other nanomechanical techniques5 besides the classic force curve method. These newer, faster imaging techniques can also measure viscoelastic response, of particular importance for polymers and biomaterials.
This article has briefly reviewed the capabilities of today's AFMs for nanoscale surface characterization. Recent instrumentation advances such as higher spatial resolution, faster imaging rates, and enhanced measurements of physical properties make AFMs more valuable than ever before. Future refinements that extend these capabilities even further will help AFMs keep pace with technology's continuing demands for better device control on smaller length scales.