Capturing Complete Quantitative AFM Nanoscale Property Data

Interview conducted by Stuart MilneFeb 9 2016

Thomas Mueller, Director of Development Applications in the AFM business unit of Bruker’s Nano Surfaces Division, talks to AZoNano about Bruker's Dimension FastScan^® AFM and the benefits of using bandwidth for faster, real time atomic force microscopy research.

The name FastScan suggest the system will be fast. In what ways is the FastScan “fast”, and what advantages does “fast” really provide users?

In a nutshell, with the bandwidth of the FastScan system one does not just get more done, one also has new opportunities for dynamics studies. For instance, the Hobbs group in Sheffield has studied polymer phase transitions. With high-speed imaging on FastScan one can observe such processes as lamellar growth rates during crystallization. Using this technology the group discovered that these rates vary in a manner not predicted by conventional models.

It turns out, there is a lot more too fast AFM than dynamics studies. There is also statistics, or one might call it “big data.” Anyone who has used a conventional AFM knows you often spend hours choosing your image site carefully because it takes a long time to get a single good image. FastScan is a tip scanner, so scan speed is completely independent of sample size and mass. And the stage is motorized and programmable. One can set up multi-site imaging at frame-per-second rates. Whether it’s a molecular binding event in liquid or a polymer phase transition, one can easily obtain hundreds of data sets. For the first time, the big data approach is feasible with an AFM.

High-speed imaging capturing the emergence of anisotropy in the phase transition of mesomorphic polydiethylsiloxane.

FastScan is listed as a large-sample system that doesn’t sacrifice high resolution. What does this mean, and how does it accomplish both?

It used to be conventional wisdom that you need a small-sample AFM to obtain the highest resolution. Such as atomic defect resolution on crystals, molecular or submolecular defects on soft matter. We saw a great benefit in removing this limitation. There are so many samples that do not fit on a 15mm puck and still where highest resolution is still of great interest. So in developing the Dimension FastScan platform, we set out to create a highly stable mechanical loop. But the most interesting advance we brought to the platform is in the area of force control and feedback through PeakForce Tapping^®.

To achieve high resolution with an AFM, the force on the probe needs to come from a small area. It turns out, how wide an area the probe ‘feels’ depends highly on tip-sample distance. Closer is better, until you start distorting or damaging the surface. Basically, you want to be close but not too close. With conventional, resonant TappingMode, forces felt during the entire amplitude of the oscillation cycle contribute to the signal. This makes it hard to get to the ultimate resolution. With PeakForce Tapping, we take an entirely different approach. It’s all about the instantaneous force detection. Only the force from the endpoint of the cycle is used for feedback, and it is detected very sensitively, to the pico-Newton level, irrespective of adhesion. As it turns out, this is the ideal way to use the principle of ‘close but not too close’. And so, whether it’s the double helix structure of DNA, major and minor groove, as shown by Pyne (University College London) and Hoogenboom (London Centre for Nanotechnology), or atomic defects, FastScan is the ideal AFM platform.

One can find amazing AFM high-resolution examples in the product marketing as well as scientific literature, but the traditional thought is that these are the result of considerable AFM expertise and careful preparation on part of the researchers. When a customer who is relatively new to AFM runs their sample with FastScan for the first time, what is the realistic spatial resolution they should expect?

With AFM of course sample preparation is always important. That said, one of the unique advantages of PeakForce Tapping is the breadth of samples, where one can achieve highest resolution. The principle I mentioned of being ‘close but not too close’ is particularly important on soft samples. Often the range of ideal tip-sample forces and distances can be very narrow, making high-resolution imaging particularly elusive.

For example, with PeakForce Tapping we have seen submolecular resolution in a thin film of isotactic PMMA, and have even resolved a molecular defect on a crystal of a polydiacetylene (PDA) compound, in air. And in both cases, not just in topography, but also in nanomechanical stiffness and adhesion maps. This also translates to electrical measurements. For example, Desbrief et al have published images of stereoregular P3HT, resolving individual molecular fibers in a conductivity image with PeakForce TUNA^®. We see high resolution where it was just not possible to achieve before.

It seems AFMs can still be complex to operate. How easy is it to operate FastScan?

The hardest issue for getting good images in AFM is not that you might drop a probe along the way. The hardest issue is optimizing the feedback loop. Not only that, there is a PI loop to optimize. In conventional TappingMode, as one is trying to dance on top of a resonance, the behavior of that feedback loop is highly nonlinear. On FastScan with PeakForce Tapping, the feedback operates off-resonance, and therefore is linear. That not only makes it more well behaved to begin with, it also allows us to implement the truly general auto-optimization schemes we have developed, our ScanAsyst^® technique. With ScanAsyst, one can obtain optimized images without any prior sample knowledge. So that then lends itself even to pre-programming a series of measurements on a range of different samples at once. And with the bandwidth and the large motorized stage of FastScan, that is exactly what one can do.

How have you achieved flexibility and ease of use and does one impact the other or vice a versa?

One thing that makes atomic force microscopy so interesting is how vibrant the field is with new and innovative measurements. Interestingly, the Dimension FastScan and Dimension Icon^® platforms stand out in peer-reviewed publications discussing the invention of new modes or approaches. This is the case for electrical modes, where for example Pierre Eyben at IMEC has invented FFT-SSRM. Or in mechanical modes, one might highlight the nanoscale DMA work of Sokolov et al. Why is that? It’s not just that all the automation, such as ScanAsyst, can be turned off and that dozens of advanced modes are available. The same elements that contribute to ease of use also make the system flexible, for example the motorized stage. Overall, the platform provides a unique combination of software access, signal access, and a wide open stage, making it an ideal starting point for custom experiments.

What are the latest innovations and features incorporated in FastScan?

The FastScan platform is very much at the center of our new technology development. It’s not just the design and implementation of a platform, to perform high-bandwidth atomic force microscopy. It has also been in our focus to enable measurements that were previously not possible. For example, the development of a way to make electrochemical measurements in Lithium ion battery research. We use the force control of PeakForce Tapping to study the fragile SEI layer that is so critical to battery performance and lifetime. Once we could do this, we worked on developing an electrochemical cell and environmental control at the level of 1ppm oxygen and water. Another example is performing high-resolution conductivity imaging on soft samples such as organic photovoltaics. The longstanding dream of performing conductivity measurements without resorting to contact mode and its destructive lateral forces is now a reality. We have seen amazing resonance in the community to these new capabilities. For example, there are more than 100 peer-reviewed publications on just  the PeakForce TUNA mode alone. We are very dedicated to continuing to focus on developing truly new capabilities and possibilities for AFM technology.

You have mentioned a range of modes. In going beyond topography with AFM, how well can one understand or quantify those other signals?

At Bruker, we consider one of the most important issues for atomic force microscopy is to not just go beyond topography, but to truly deliver quantitative data. This is a challenge for AFM technology. TappingMode has been the go-to mode for many years, yet it is especially challenged when it comes to quantifiable nanomechanical data. To begin with, you measure only a few parameters in TappingMode, just amplitude and phase. Even if you measure a second Eigenmode, it is still only a total of four parameters measured. Not only is that a very few compared to the hundreds of data points available in a force curve, but it is also lacking when compared to the sample parameters that need to be determined. The influence of several properties is necessarily convolved and very hard to separate. For example, adhesion contributes to the resonant phase, and this is usually neglected. So in TappingMode it is not just challenging to quantify the data precisely, it is even challenging to just separate the contributions to the signal and attribute contrast unambiguously. This is why we found it so important to develop a powerful imaging mode based on force curves. And this is precisely what PeakForce Tapping does for the first time. It creates an imaging mode that is fast and highest resolution, based on force curves. Force curves are like nano-indentations, so you know you can quantify that. You still need probes with accurately known spring constants and end radii, but starting with force curves in the first place is the right approach.

Most stable tip-scanning AFM.

Who are the typical customers for this system?

Many customers use the FastScan platform in advanced materials research. For example, the research group of nobel laureates, Geim and Novoselov, at the UK’s National Graphene Center in Manchester recently published a paper where PeakForce QNM studies revealed a subtle stress-strain relief effect in graphene on boron nitride, to help understand how to build 2D heterostacks. And numerous groups have used FastScan for high-resolution property mapping on advanced functional materials, such as photovoltaics. Due to its ability to produce highly accurate data with high throughput, FastScan also finds industrial applications in data storage, as well as in the chemical and pharmaceutical industries.

How open is the system to modifications by your customers to do new kinds of experiments?

One of the great strengths of FastScan lies in the opportunities it affords for performing new measurements that have not been done before. For one thing there are the opportunities for new measurements that arise from our rapid development of unique modes.

The publication record bears this out, for example, with more than 1000 publications now referencing PeakForce Tapping use. This is very gratifying for us at Bruker, to have developed these modes that are changing the way the community does research. And then there is the flexibility of the system, which allows researchers to further develop atomic force microscopy as a technique and literally invent new modes. We find this fascinating and design our systems to maximize the opportunities for it. There are three elements you need. First, there has to be external software access or scripting. Second, users need direct access to raw signals, such as deflection, to manipulate feedback or implement new detection or processing schemes. And third, a wide open stage is required for placement of home-built accessories or actuators. What sets FastScan apart in terms of potential for new experiments is that it supports all three of these elements so well.

About Thomas Mueller

Dr. Thomas Mueller is the Director of Development Applications in the AFM business unit of Bruker’s Nano Surfaces Division. Thomas has been with Bruker for 12 years having held positions in applications and product management, and is the author of over 50 publications, reviews, and application notes. He received his Ph.D. in 2000 from Yale University on developing new linear and nonlinear spectroscopic probes of molecular structure and dynamics, followed by postdoctoral research at Columbia University focusing on scanning probe microscopy as a tool for interrogating self-assembly and chemical reaction specificity.