AFM-based nanoelectrical modes are used in many different domains ranging from piezoelectric materials and semiconductors through to biology and energy research.
There are modes available to define the local resistivity, conductivity, carrier-type, carrier concentration, charge, or piezoelectric characteristics with nanometer-scale spatial resolution. These modes typically need a direct contact between the sample and the AFM tip. Traditionally, the resultant data has been in the form of a two-dimensional (2D) map, produced in contact mode, with one electrical data point for each XY location. Electrical spectra or ramps are created at a few, meticulously selected sites in the 2D map to get a better understanding of the local electrical characteristics.
This article demonstrates a novel nanoelectrical imaging technique that goes one step beyond a 2D map, producing a correlated nanomechanical data cube and comprehensive electrical data cube while operating at standard imaging speeds. An added advantage of this method is that it avoids contact mode imaging, thereby extending electrical measurements to delicate and soft samples and enhancing measurement consistency by increasing the lifetime of the AFM tip. This is also considered a general method that is relevant to a majority of the nanoelectrical modes and applications.
The more complete data set includes new data and allows functional imaging on complex samples as illustrated in examples of Li-ion cathode characterization, piezoelectric domain switching dynamics, semiconductor carrier profiling, and barrier-height mapping in heterogeneous metal oxides.
Moving Beyond Property Mapping
High-resolution (<1–5 nm) topographic images are often generated using atomic force microscopy and this is done by raster scanning a sharp tip on the surface of the sample with feedback-controlled tip-sample force interactions. AFM technology was pioneered by Bruker to concurrently capture and obtain mechanical characteristics from the interactions between the tip and sample. PeakForce Quantitative Nanomechanics (PeakForce QNM®) mode and FASTForce Volume™ (FFV) mode, for instance, offer a complete force curve for each image pixel, and the contact resonance mode offers a resonance spectrum for each image pixel. These spectra make it possible to extract and visualize mechanical characteristics like stiffness, adhesion, tan-delta, deformation, modulus, and loss and storage modulus in 2D maps.
An array of application modules is available for electrical characterization, and AFM nanoelectrode probes introduced recently for scanning electrochemical microscopy (SECM) and nanoelectrical characterization enable electrical measurements in liquid. All these capabilities incorporated into a single instrument offer investigators greater experiment reach and efficiency in regular industrial applications as well as for highly multidisciplinary studies in polymeric systems, semiconductor devices, piezoelectricity, biological characterization, and energy research.
AFM Nanoelectrical Modes
AFM nanoelectrical modes are usually operated with a fixed set of operating conditions, such as frequency, AC bias, DC bias, etc., and one electrical measurement is carried out for each XY location, producing a 2D electrical map. This is also true in PeakForce Tunneling AFM (PeakForce TUNA™) and relevant PeakForce Tapping-based electrical modes that have the potential to resolve the constraints of contact mode.
AFM-based electrical characterization, in addition to 2D maps, has traditionally included point spectroscopy, in which one of the operating conditions is swept as the AFM tip is kept immobile, for instance, in the acquisition of current-voltage (I-V) curves at a few selected separate positions or a single selected separate position. It is also possible to perform point spectra in arrays, but to date, this has stayed relatively time-intensive and has only enabled partial spatial sampling. Capacitance-voltage (C-V) spectra used for investigating the delineation of pn-junction in Si devices and I-V spectra for studying the effects of humidity on nanoscale electrochemistry in solid silver ion conductors are published application examples.
DataCube modes, available from Bruker, integrate point and imaging spectroscopy, creating a built-in three-dimensional (3D) data set. This is a multidimensional data cube that determines the reliance of a particular electrical parameter in every XY position as a function of one of the (electrical) operating conditions. A big data approach has replaced the subjective method of guessing target locations (for single-point spectroscopy), enabling higher dimensional data that can be sliced along any plane or axis, and which is favorable to key component analysis and machine learning methods for data reduction. In addition, the high dimensional electrical data sets make it possible to extract high-resolution maps of electrical characteristics that were not possible before using AFM. Images of piezoelectric switching voltages, Schottky barrier height, and flat-band voltages, amongst others, are some of the examples.
Instead of using contact mode, the tip is moved from one pixel to another in an FFV method. An electrical spectrum is obtained while inserting a hold segment into each force curve in each pixel. Moreover, the force curve at each pixel integrated into the FFV map represents a series of nanomechanical data cubes that are spatially correlated with the electrical data cubes. Considering that lateral forces are prevented by the FFV map, tip lifetime is significantly enhanced and the electrical measurements are extended to fragile samples, including nanoparticles and polymers, that otherwise cannot be inspected in contact mode. In this article, the principles of data cubes are demonstrated for scanning microwave impedance microscopy (sMIM), scanning spreading resistance microscopy (SSRM), scanning capacitance microscopy (SCM), piezoforce microscopy (PFM), and tunneling AFM (TUNA); however, it can also be extended to other electrical modes in both liquids and air.
Experimental Principle and Data Acquisition
Seamless integration of current high-performance nanoelectrical modes with the FFV method has led to the development of Bruker’s DataCube (DCUBE) modes. Within the FFV mode, the tip is maintained in a fixed XY position and, at the same time, a force-distance cycle with “dwell” segment is performed in the Z direction. This is shown by the height sensor plot and the relative force curve illustrated in Figure 1a.
Figure 1. DCUBE modes operating on FFV approach: (a) height sensor (red) and deflection error (blue open circle) plots with segments showing extend (1–2), dwell (3), and retract (4–5) cycles; (b) pattern of sample bias (red) and corresponding TUNA current (blue); (c) illustration of scan pattern during acquisition of DCUBE-mode data; and (d) five TUNA current slices from the DCUBE-TUNA results.
Bruker’s proprietary low-force trigger capability enables force ramp rates of around 300 Hz. Together with dwell times of sub-100 ms for the hold segment, excellent throughput can be achieved and regular imaging speeds can also be maintained when achieving dense arrays. The ramp rate for the “retract” and “extend” segments in Figure 1a is 60 Hz at a standard force (or deflection) trigger level of 15 nN and also with a Z movement of 80 nm. This leads to 9.6 μm/second ramp speed.
The sample bias is swept from −2 V to +2 V during the “dwell” segment (here 100 ms), relative to a ramp rate of 40 V/second, as depicted in Figure 1b. For the dwell segments and the non-dwell segments, the number of data points can be set independently, enabling high spectral resolution. In this case, 256 data points result in a resolution of 0.625 nm in the Z direction and 190 data points lead to a resolution of 21 mV in sample voltage. Increasing the amount of data points further can result in higher density. The four spectra — deflection error, sample bias, height, and TUNA current — are concurrently captured.
Following the spectral acquisition in one pixel, the tip shifts to the subsequent pixel using the FFV raster scanning technique (Figure 1c), preventing the shear force that is standard to contact mode–based scanning. Figure 1d depicts five standard current slices obtained from a data cube achieved with the help of this principle on a γ-Fe2O3 sample. Every current slice image contains 256 x 256 pixels for a 3 x 3 μm2 sample area; 190 of these current slice images are provided by the 190 spectral data points from the dwell segment. Simultaneously, the “extend”/“retract” segments of the deflection plot build an entire force-distance curve, making it possible to obtain quantitative nanomechanical images for stiffness, modulus, deformation, and adhesion.
In DCUBE modes, the conditions for the electrical dwell period and the force curve cycle are individually controlled, enabling users to independently refine the measurement of electrical spectra and force-distance spectra. The nanoelectrical mode that is being used decides the period of the hold segment for superior quality signals. To be more specific, it relies on the applied AC frequency and/or the sensor’s bandwidth. Since the bandwidth of sensors is usually in the kHz range (for instance, the typical bandwidth of the PeakForce TUNA application module is 15 kHz), acquisition times for capturing the entire electrical spectral down to millisecond level can be realized.
It must be noted that when the electrical ramp speed is increased, the signal/noise ratio will be affected. For instance, a series of capacitance-voltage (C-V) spectra of a p-type semiconductor captured by DCUBE-sMIM is shown in Figure 2. These are actually raw data without averaging. For every spectrum, the sample bias ramp rate and the dwell time are shown in the figure.
Figure 2. A series of C-V spectra of a p-type semiconductor captured by DCUBE-sMIM. These are raw data without averaging. The dwell time and the sample bias ramp rate for each spectrum are indicated.
As demonstrated by these results, a ramp rate of 80 V/s or a dwell time of 50 ms is more than enough for capturing all the sample’s spectral features. With regards to both force and electrical spectra, the time for each pixel is usually in the range of 20–200 ms. For a pixel map of 256 x 256, this relates to an acquisition time of 22–220 minutes. Although, with this comparatively long period of spectroscopic mapping, over 1000 data points can be captured for every electrical spectrum, leading to over 1000 dynamic slice or sequence images at a capture rate of 1.3–13 seconds per image.
As an alternative to ramping the DC sample bias at the time of the dwell period, users can even choose to ramp any kind of parameter pertinent to the chosen electrical mode (for example, the frequency, amplitude, or AC voltage). In this manner, a wide range of electrical spectra, such as C-V, R-V, I-V, dR/dV-V, dC/dV-V, PFM-amplitude-frequency spectra, and PFM-amplitude-V can be collected. The available electrical DCUBE modes are summarized in Table 1.
Table 1. Summary of DCUBE modes and electrical characterizations with FFV
||Force Stiffness Adhesion Modulus
In addition, the electrical conditions can be kept constant during the dwell time, which can be long (tens of seconds) or short (milliseconds). In this fashion, it is possible to average the electrical signal over a user-defined period, leading to enhanced signal/noise ratios when compared to the equivalent PeakForce electrical mode or contact mode. The electrical parameter is determined as a function of time during the dwell period, paving the way for analyzing time-dependent phenomena. Examples include tracking time-dependent dielectric breakdown in dielectrics; mapping of time-dependent photoconductivity or conductivity following an external stimulus, like an electrical pulse or light pulse; analyzing charging or discharging effects following the application of a voltage pulse, etc.
Through the DCUBE modes, nanometer-scale mechanical and electrical properties in high-density data cubes can be captured simultaneously. This overcomes long-standing characterization and efficiency obstacles for materials scientists and engineers. By processing such information-rich, big data results, creative data mining can be stimulated and encouraged. The standard NanoScope® analysis offline software from Bruker offers a suite of data tools for simple processing and visualization, for example, the extraction of separate spectra and slices. Some of the main functions for handling DCUBE data (here, applied to DCUBE-TUNA data obtained on a γ-Fe2O3 sample) are shown in Figures 3 and 4.
Figure 3. Processing, analysis, and visualization of DCUBE data. Slices of the spectroscopic maps can be extracted for regular processing of AFM images in NanoScope Analysis software: (a) surface topography; (b) quantitative nanomechanical properties—adhesion map; (c) a slice map from TUNA current—sample bias (I-V) spectra at −1 V (cover image of a movie created from 130 slices available on the Bruker website); (d) slices of TUNA current maps at selected voltages; and (e) scanning barrier height images created by analyzing all I-V spectra. The color bar displays from ohmic (zero) to insulating (one). NanoScope Analysis software goes with a MATLAB toolbox that allows one to extract the whole cube of data for further analysis in MATLAB. Image (e) was created by MATLAB.
Figure 4. Processing, analysis, and visualization of nanoelectrical data from DCUBE-TUNA mode with NanoScope Analysis software: (a) current slice at +1 V, (b) point-selected I-V spectra as numbered in (a), (c) density plots based on 1890 I-V spectra from the dashed-yellow-square region in (a), and (d) TUNA current changes over sample bias across the dashed-yellow line in (a) where the white plot on the image is the corresponding surface profile with a range of 150 nm in height. This is a two-dimensional slice of the data cube at a fixed Y position as indicated by the coordinates on the image.
Firstly, as one would do with standard AFM images, slices of the spectroscopic maps can be obtained for additional processing. This enables the extraction of surface topography and quantitative nanomechanical maps, as illustrated in Figure 3a and 3b. A current slice at a fixed sample bias of −1 V is shown in Figure 3c, while more current slices at nine chosen voltages are depicted in Figure 3d. A current slice at +2 V collected in a different area on the γ-Fe2O3 sample is shown in Figure 4a.
Secondly, any of the spectra can be selected, displayed, and exported through single- or multiple-point selection in the nanoelectrical, nanomechanical, or surface topography images. All the spectra from a chosen region can also be displayed and overlaid. In addition, a dense collection of spectra like this can be visualized through statistical techniques with the integrated density plots.
An example is illustrated in Figure 4c, produced from 1890 spectra in the dashed-yellow-square region indicated in Figure 4a. The darkness in this plot matches with the spectrum’s population density in this target region. In this specific case, the three apparent curves shown in Figure 4c indicate three clear electrical domain behaviors — high barrier, low barrier, and insulating.
Thirdly, the advanced NanoScope Analysis software also offers electrical contour plots along a line defined by the user. Figure 4d illustrates the TUNA current variations over sample bias across the dashed-yellow line indicated in Figure 4a. This is actually a 2D slice of the data cube at a fixed Y position, as specified by the coordinates on the image. Every vertical line includes spectral data rendered by the false color of TUNA current. The location along the line in Figure 4a is indicated by the (dark) blue-to-yellow stripes and where TUNA current alters with sample bias. The white plot shown in Figure 4d is the related surface profile along the dashed-yellow line indicated in Figure 4a, with a range of 150 nm in height. This demonstrates that the nanoelectrical characteristics do not clearly correlate with the surface topography and only arbitrarily differ among various particles.
Additional detailed analysis and data handling can also be carried out using specialized external analysis packages, like Python- and MATLAB-based tools. A toolbox is offered for this purpose, enabling users to directly import the DCUBE raw data files and also extract the target channels. After converting these raw data files, they can even be fed to other programming languages, like Python. For instance, movies can be created to demonstrate the materials’ electrodynamics. Figure 3c shows the TUNA current map, which is the cover image of a movie produced from 256 slices obtained from the data cube. For this kind of measurement, the sample bias ramping is from −1 V to +1 V.
In a second instance, a batch analysis of all the I-V spectra in this data cube was done using MATLAB: The barrier height of the metal tip/γ-Fe2O3 semiconductor Shockley junction was extracted by fitting the I-V spectra with a thermionic emission model. Figure 3e shows the ensuing barrier height map, which demonstrates heterogeneity among varied γ-Fe2O3 nanoparticles. The color bar indicates from ohmic (zero) to insulating (one), confirming earlier studies by sMIM and PeakForce TUNA demonstrating clear electrical characteristics at the particle level. This is a standard case of how the huge amounts of data in combination with detailed analysis make it possible to obtain new, quantitative, nanoelectrical information that was not possible before by using traditional AFM modes.
This article is the first of a three-part series on AFM-based nanoelectrical modes:
- Pittenger, B., Erina, N. and Su, C., “Quantitative Mechanical Property Mapping at the Nanoscale with PeakForce QNM,” Bruker Application Note 128 (2012).
- Pittenger, B. and Yablon, D. G., “Improving the Accuracy of Nanomechanical Measurements with Force-Curve-Based AFM Techniques,” Bruker Application Note 149 (2017).
- Berquand, A., “Quantitative Imaging of Living Biological Samples by PeakForce QNM Atomic Force Microscopy,” Bruker Application Note 135 (2011).
- Kaemmer, S. B., “Introduction to Bruker’s ScanAsyst and PeakForce Tapping AFM Technology,” Bruker Application Note 133 (2011).
- Pittenger, B. and Yablon, D. G., “Quantitative Measurements of Elastic and Viscoelastic Properties with FASTForce Volume CR,” Bruker Application Note 148 (2017).
- Li, C., Minne, S., Pittenger, B. and Mednick, A., “Simultaneous Electrical and Mechanical Property Mapping at the Nanoscale with PeakForce TUNA.” Bruker Application Note 132 (2011).
- Li, C. et al., “PeakForce Kelvin Probe Force Microscopy,” Bruker Application Note 140 (2013).
- Huang, Z. et al., “Nanoscale Mapping of Permittivity and Conductivity with Scanning Microwave Impedance Microscopy,” Bruker Application Note 145 (2016).
- Calahorra, Y., Smith, M., Datta, A., Benisty, H. and Kar-Narayan, S., “Mapping Piezoelectric Response in Nanomaterials Using a Dedicated Non-Destructive Scanning Probe Technique,” Nanoscale 9, 19290-97 (2017).
- Huang, Z. et al., “An Introduction to AFM-Based Scanning Electrochemical Microscopy: PeakForce SECM,” Bruker Application Note 147 (2017).
- Jiang, J. J. et al., “Nanoelectrical and Nanoelectrochemical Imaging of Pt/p-Si and Pt/p+-Si Electrodes,” Chemsuschem 10, 4657-63 (2017).
- Nellist, M. R. et al., “Atomic Force Microscopy with Nanoelectrode Tips for High Resolution Electrochemical, Nanoadhesion and Nanoelectrical Imaging,” Nanotechnology 28, 095711 (2017).
- Nellist, M. R. et al., “Potential-Sensing Electrochemical Atomic Force Microscopy for in operando Analysis of Water-Splitting Catalysts and Interfaces. Nat Energy 3, 46-52 (2018).
- Toma, F. M., “Disentangling Interfacial Energetics,” Nat Energy 3, 6-7 (2018).
- Edwards, H. et al., “Scanning Capacitance Spectroscopy: An Analytical Technique for pn-Junction Delineation in Si Devices,” Appl Phys Lett 72, 698-700 (1998).
- Yang, S. M. et al., “Humidity Effect on Nanoscale Electrochemistry in Solid Silver Ion Conductors and the Dual Nature of Its Locality,” Nano Lett 15, 1062-69 (2015).
- De Wolf, P. et al., “Data Cube Modes: Enabling Functional Imaging with Higher Dimensional Electrical Data Sets,” Microscopy Today (2018).
- Bruker NanoElectrical Lab, https://www.bruker.com/products/surface-and-dimensional-analysis/atomic-force-microscopes/nanoelectrical-lab.html (2019).
- Ramachandran, P. and Varoquaux, G., “Mayavi: 3D Visualization of Scientific Data,” Comput Sci Eng 13, 40-50 (2011).
This information has been sourced, reviewed and adapted from materials provided by Bruker Nano Surfaces.
For more information on this source, please visit Bruker Nano Surfaces.