Forums
Bruker Media
Community Media
Bruker AFM Probes
SPM Digest
Application Notes
NanoTheater
Website
中文
Brochures & Datasheets
Publications
Probes Catalog
Events
Manuals & Documentation
Presentations
Guide to AFM Modes
News
Journal Club
Webinars & Video
Nanovations
Other
Achieving atomic resolution with AFM has been a goal of the community since the invention of the technique. There are a number of great results from academic laboratories (Fukuma et. al., Hoogenboom et. al., Gross et. al.) that have demonstrated this by systematically lowering the noise of their systems - ie moving to restrictive small sample platforms, small cantilevers, etc. At Bruker, the challenge we made to ourselves was: can we achieve atomic resolution with a highly productive large sample platform using standard cantilevers. The focus was to develop an approach with no compromises - Fastest, Highest Resolution, Large Sample, All Modes and Accessories.
See the image of the cleavage plane of calcite taken with the Dimension FastScan using a standard cantilever (SNL+, 60um) in water. In the image you can see two crystal planes (brown and blue) separated by the dissolving crystal front (red). We have overlaid the atomic model of the oxygen atoms in the calcite on the lower plane, and there you can see the zig-zag structure revealed in the image. You can also see the individual defects in the crystal near the dissolution plane as well as a secondary facet brought out on the etch front. Many who have seen this have commented that it is remarkable that this is possible on a large sample system with a regular cantilever. In addition to the fundamental performance of the FastScan, this fidelity is enabled through Peak Force Tapping’s ability to control instantaneous tip-sample force. We will be announcing a webinar shortly that will explain this in detail. In the meantime enjoy the image!Special thanks to Bede Pittenger & Yan Hu of Bruker Nano, and Daniel Ebeling of the University of Maryland, who acquired this data in the Santa Barbara Laboratories of Bruker’s Nano Surfaces Division.Steve
I had a comment on another forum that I thought I would repeat here about our work with respect to that of Professor Frieder Mugele at Twente University. Professor Mugele has taken amazing atomic images and force curves on Gibbsite using the MultiMode. He and Dan Ebling have a paper in process, which should be out soon. We are also writing a joint Application Note with them - I will post links to both when available.That work was an inspiration to our effort and highlighted the benefit of getting Mechanical property maps simultaneously with imaging. In our experiments, are able to obtain all of the force curve analysis channels of Peak Force QNM at atomic resolution.Here is a link to an atomic resolution image of Calcite, this time taken on the Dimension Icon (90um scanner, large sample platform, SNL+ probe) showing the height and stiffness channels. In the software, this is called Modulus, however, at the atom level, modulus is not defined (since it is a continuum property). Here we simply note that areas where the force curve slope is steeper appear brighter in the image, and are qualitatively stiffer than other areas. In the image you can see striking structure including the very interesting detail that the alternate rows of atoms have significantly increased stiffness.Again, note that this was taken on a large sample system compatible with a complete suite of modes and accessories (including the Dimension FastScan, PF-TUNA, Heating/Cooling. . .), and with a regular cantilever. SteveBruker AFM
Fantastic images from Physics of Complex Fluids.
*** Copied from LinkedIn Posting ***
I was wondering could it be possible that conventional tapping mode in liquid can achieve similar resolution. Since you are using a standard cantilever, the overall dynamics didn't change, so you lowered the noises at all levels? What have changed so that you can have higher resolution than tapping mode, stiffness information aside? The reason I am asking is that I've seen in the other thread started a few month ago, they used the same sample and could observe the atomic defect on the calcite surface in liquid with very small amplitude (In Manager's Choice in case you didn't see it). Don't get me wrong, the images are impressive and I am not judging just curious. Maybe you could elaborate more.
Yes, indeed in addition to PeakForce Tapping, we have also demonstrated atomic resolution in TappingMode on FastScan, Icon, and MultiMode. The image below show a Tapping example on Mica using the MultiMode 8 and a SNL+ cantilever (60um long). In all these cases, we did not redesign any of the systems, and used them in the standard configurations - but taking care to locate them in quiet parts of our building.
Your question brings up an important point about achieving this kind of resolution. A quick “order of magnitude” calculation highlights that, from a noise perspective, you don’t need small levers to achieve atomic resolution: for example if you have a regular lever with ~100fm/rtHz, imaging in a normal bandwidth, say 1k, the thermal noise will only be 3pm. On Calcite, for example,we measure atomic heights of ~50pm so this is not the limiting factor in terms of noise.The reason why many Tapping approaches use small levers is that they help to enable Tapping at ultra low amplitudes - but this is the key point, why go to ultra low amplitudes? It is to mitigate the fundamental problem that Tapping integrates all force interactions over the entire Tapping cycle.The cycle averaging of Tapping limits performance because 1) the high resolution tip-sample interaction only occurs when the tip is close to the sample, and this is a fraction of the cycle, and 2) at low imaging forces, the effects of long range forces dominate the cycle. (This is also why atomic Tapping images are done in fluid. . . eliminate the long range capillary force.) A better approach is to use PeakForce Tapping because it does *not* measure the average force over the entire cycle, but only the desired high resolution repulsive interaction when the tip is closest to the sample. We will have a webinar shortly that describes this in detail, but you can see start to see this in the slide below.
There is a lot on the graphic, but the key take away is the force vs time trace on the bottom right. Tapping averages all of the attractive forces (blue) and repulsive forces (red); making it hard to tease out the desired high resolution interaction (red). PeakForce Tapping, lets you measure the peak repulsive force (red) independently, and it is why we can get atomic resolution any of our systems, without any limitations or compromise in terms of sample size, AFM flexibility (modes & accessories), or costly/rare probes. Steve
I posted this answer to a question on another forum, and thought I would repeat it here. You can see the question at: http://goo.gl/yLoCF***Thanks for your comments, and we, along with several others who posted in these forums, are in agreement that atomic resolution imaging using Tapping in fluid is not new, despite recent attempts to package it as such. This was not our interest pursuing these experiments. Atomic resolution is a good benchmark, and being able to routinely achieve it on our instruments helps us to continue to design the world’s best AFMs. Indeed, the Dimension 5000 was a great instrument, and its successor of several generations, the Dimension FastScan is even better. In addition to the singe atom resolution images shown above, its dynamic imaging capabilities are unrivaled (see some example experiments at http://www.youtube.com/user/BrukerNano, and posts in these forums); additionally because of the large sample format, and the FastScan’s full suite of modes and accessories, there is no compromise that has to be made to get this speed and performance. As I described in the previous post, PeakForce Tapping provides a better way to achieve high resolution because you measure instantaneous force rather than cycle average force. Another great advantage of PeakForce Tapping, particularly when imaging at the atomic scale, is you can obtain a forces curve for any pixel in the image.
Here are two examples of force distance curves collected during Peak Force Tapping imaging. One set of approach / retract curves was collected on calcite and the other was collected on mica. The mica data shows a long-range repulsion which suggests the presence of an electrostatic double-layer or hydration layers of water molecules. Additional experiments are required to further understand the source of this long-range repulsion, and its variation over individual atoms in the lattice. We are presenting it here only as an example of the capability.It should be noted that, unlike Tapping, this spectroscopic capability is now accomplished simultaneous with imaging, thereby increasing the accuracy and confidence of the localization, as well as the speed and ease of the measurement. With PeakForce Tapping there is no ambiguity about which force curve corresponds to which atom. To your question, since modulus is a continuum property and not defined at the single atom level, we providing the force curve, or stiffness map, but don’t convert to modulus. There are numerous examples in the literature where researchers are obtaining quantitative modulus data, at the appropriate scales, using PeakForce QNM. When you mentioned force control to us previously (at MRS I believe), you were interested in seeing PDES imaged with PeakForce Tapping. We did that a while back, and I thought you saw it. I can post either data set.Best,SteveBrukerAFM
A follow up to the previous conversation, from another forum:Thanks for bringing the references of your high resolution work into the discussion. They are as impressive now as they were then. I do have to say I am pleased that all of those experiments were completed on Bruker AFMs, particularly the Dimension Line. It is also nice to show that our systems have been accomplishing atomic resolution in TappingMode mode for over the past decade. While we are excited about the benefits of PeakForce Tapping, that does not change that we invented TappingMode and have been doing it ever since. With a Bruker system, it is not a choice between Tapping and PeakForce Tapping; you always get both - there is no compromise that has to be made.Regarding:http://dx.doi.org/10.1063/1.1697629 (APL 2004, 84, 2697)http://jjap.jsap.jp/link?JJAP/45/2158/ (JJAP 2006, 45, 2158-2165)and, if I interpret your comments correctly with respect to these papers, I believe we are in agreement: Tip shape has a powerful influence over what is resolved in the image, even if a defect is visible (ie Figure 9, JJAP), and in fact, it may even be easier to see a defect than the lattice itself (Figure 9c). That said, and including in PeakForce Tapping, we have all seen defects, and all our lattices look similar, but in none of these cases does this absolutely imply there are not multiple tip sample interactions. Continuing to follow this line starts to transition the discussion into a “tip-sharpness” discussion rather than a “system performance” discussion. Both are valuable, but my initial intent was to focus on the system performance aspect. That said I can’t help but note that you used a specialized tip, on a 225um lever, in your APL paper and were rewarded (as contemplated by the JJAP reference) with very large corrugation - up to 600pm. Our samples are completely different, but compared to other’s work on similar samples, we saw large corrugation (~50-200pm) using a standard probe. The large corrugation is good evidence of resolution. Again, we are not the first to look at this sample, and we are seeing what others have seen. The difference, in my opinion, is we are doing so in a better way: on a large sample, high speed, fully accessorized, unmodified, commercial AFM using a standard probe. Additionally, we are able to simultaneously acquire force measurements at any pixel in the image, and this is a big advantage - as demonstrated on both calcite and mica. http://goo.gl/Xn4OxThe other reason I mention these numbers is to put the detector noise into context. In both of our cases, our corrugation is much larger than the detector noise contribution: consider a system has deflection noise as high as 100fm/sqrt(Hz), for a normal 1kHz imaging bandwidth, the noise is 100fm*sqrt(1000) = 3.2pm, which is still much less than the ~100pm of corrugations shown on various atomic resolution images acquired by you, by us, and by multiple other groups (and why you were able to do it with a 225um lever).It is also good to note detector noise is not an independent metric, but scales with optical sensitivity, which scales with probe length. While you did not in your request, quite often this number is quoted without a probe length specified, which can be confusing. In our experiments we typically used the SNL probe (60um), and measured noises of ~25fm/rtHz. I expect in your APL experiment, with a 225um probe, your detector noise was ~200fm/rtHz - again this is not the limiting factor for resolution.On the soft imaging, I asked around the team here, and I don’t think we ever got back to you with the data I mentioned in my previous note. I will put that together and reply later.Best,SteveBruker AFM
Hi Steve,
Great images indeed! We use PeakForce Tapping quite a lot here in Nottingham, LBSA. And we do get some quite remarkable images as well.
Although we mostly use it for soft materials like organic monolayers.
Here is the link to our recent paper with some high resolution images aquired with PeakForce:
http://pubs.acs.org/doi/abs/10.1021/jp212388c
Kind regards
Vladimir
We previously posted a link showing atomic defects in the Mica lattice taken in TappingMode (http://goo.gl/wvVFi). Indeed this is routine for us to do on our different platforms, on multiple samples. If you are interested in these types of experiments, please contact me offline so I can arrange for you to see them at any time.In looking at the 4 previously posted PeakForce Tapping images, it should also be clear that we can also do this in PeakForce Tapping on multiple platforms, and multiple samples. I am posting another sequence that highlights unique lattice defects, as well as adsorbates. In this sequence you can also see 2 different adatoms appear on the surface and then disappear in subsequent frames. Likewise, we are more than happy to demonstrate these experiments at any time.
To me, these demonstrations of system performance are not about experimental specifics (ie Have you captured an adsorbate appearing and disappearing yet?), but how the instrument will make you more productive in your research - in my opinion PeakForce Tapping provides a better way: 1) Look at the images in the postings, and from other sources, to compare noise between Tapping and PeakForce Tapping. The PeakForce Tapping images are certainly not “several times noisier”, and in our experiments, they are often better than those obtained in Tapping. This highlights a significant shift in the high resolution paradigm, and more importantly, the thinking about what the critical noise sources are. Up to now, ultra-low amplitude tapping has been the primary approach, but it has its limitations (must be done in fluid because the small amplitudes cannot overcome surface adhesion, it leads to restrictive instrumentation designs, it cannot distinguish attractive and repulsive forces in the cycle, it may induce tapping force variations on heterogeneous samples, etc.). By eliminating the background force contributions from TappingMode’s cycle averaging, and moving to the instantaneous force control of PeakForce Tapping, these limitations are overcome, and this opens up a wider range of applications and experiments as well as new opportunities for scientific research. 2) Regardless of one’s bias, it has to be acknowledged that taking force measurements synchronously with imaging, particularly at high resolution, is a better way than: imaging, changing modes, taking curves, switching back, and making sure you have not drifted from your area of interest. The need for “creep corrections” in image sequences more than reveal the need for this kind of improvement. While indeed, these measurements are not the first of their kind, the simultaneous acquisition of force curves with imaging at atomic resolution is, it represents a better way - and it is now readily accessible (http://goo.gl/Xn4Ox) 3) That said, why bother to take force curves only at a few points, when you can obtain, and analyze them in real time, for every point of an image. (http://goo.gl/gw9Ix) PeakForce Tapping lets you immediately move beyond topographic imaging, so you can learn more about your sample without sacrificing anything. At the same time, and as shown by the images in the posts of others, our AFMs lead in TappingMode imaging, and with our systems you always have the option of low amplitude tapping in addition to PeakForce Tapping.Our June webinar will explain this in much more detail. Its coordinates will be posted soon. I would also like to again thank Daniel Ebeling of the University of Maryland (especially on the Mica), and Bede Pittenger & Yan Hu of Bruker Nano, who acquired this data in our Santa Barbara Labs.SteveBruker AFM
The atomic PeakForce capture function seems somewhat different to standard PeakForce tapping, where (as far as I'm aware) a force curve is performed at each pixel, but not recorded.
Is there a way to retain the force curves at each pixel in any PeakForce tapping or ScanAsyst image?
Cheers,
Ash
*** Following on from a post in another group ***. . .I am also a bit remiss in posting the “PeakForce Tapping: Atomic Imaging” webinar link. It is first thing tomorrow morning so please register now - Join us at: http://goo.gl/z3l6cWe will be showing more data than we have posted here - including getting atomic resolution IN AIR, as well as repeating the PDA sample with very interesting results. The following work I will show is on PDES which was carried out by Natalia back in December of 2011 (http://goo.gl/olNkq). For reference I have also included some of your work on the same sample (http://goo.gl/iETDf).In your work, you used two set points in TappingMode to highlight how one could “skim” the surface of a highly dissipative sample, or by moving to harder tapping, causing changes in the phase image to reveal the underlying surface (but unable to materially affect the height image).Because we are not dealing with resonant dynamics in PeakForce Tapping we have the ability to do the same, but also more, in that we can easily control the interaction force and watch its effect on the height image. The direct force control lets us either a) image the highly dissipative top surface, or b) press thorough this layer and imaging the underlying structure - all with the same cantilever. In our data we are imaging the exact same location using the exact same cantilever (0.42 N/m) and we can get both the low force and high force characteristics in PFT (bottom) - whereas in TappingMode (top), we are only able to influence the (difficult to interpret) phase contrast (not shown), while the height images remain constant.The line traces examine the impact that force plays on the apparent height of this sample. Being able to explore the entire force range, in a single experiment, is another example of how Peak Force Tapping can help you to better understand your sample.We consider (and as I heard in a separate conversation at MRS 2011 also from you) PDES is equal if not superior to SBS for this type of examination.Best,Steve
In the released NanoScope version today it is possible to capture the force curves for a partial image using our High Speed Data Capture function and correlate the position of those force curves with the image pixels.As the name perhaps hints, PeakForce Capture will become a more powerful feature soon. A release later this summer will extend the capability to capture the force curve for each pixel in entire images. The feature is still under development, so additional details will need to wait for the release.
*** Following on from a post in another group ***. . .
Thanks for your review. I believe we agree on the assessment of the data, and in several of its interpretations, but disagree on some key conclusions. Where we agree is PeakForce Tapping is a mode that provides unique and complimentary information to existing modes. As I have mentioned before, with the Bruker platforms this is only a benefit, because in addition to PeakForce Tapping, all of the standard modes are available on all our systems - so no compromise needed in choosing this technology.Where we differ is in the value of PeakForce Tapping’s (PFT) ability to do the things you say are difficult for TappingMode, and in the implication that Tapping Mode is gentler than PFT. To these points, it should be noted that the feedback loop in Tapping mode does not control the force between the tip and sample – it controls amplitude variations which arise from both change in dissipation from the energy stored in the cantilever’s vibration, and from frequency shifts of the cantilever resonance. Because TappingMode is not directly controlling tip-sample force, there are often large variations across the sample- particularly when scanning samples where there is large roughness (e.g. deep, narrow trenches) or there are material property variations (e.g. PDES on Si). These variations can result in hundreds of nN of maximum force. In contrast, with calibrated PFT, we know the force everywhere on the sample, and can control it down to the level of tens of piconewtons.As a baseline, we are able to achieve force control, and images, equal to or better than can be achieved in Tapping mode (http://goo.gl/olNkq, Left Top - Light Tapping, Left Bottom - Light PFT). It is true that we use a different cantilever, with a different spring constant, but that is because we use a direct method of force control and do not rely on Tapping’s resonant dynamics (via Q) to reduce the interaction force produced by high spring constant cantilevers. This is the paradigm shift of PFT which I believe is so important - interact with your samples at the force you choose by using cantilevers of appropriate stiffness *rather* than use very stiff cantilevers but rely on the uncertain average resonant interaction to bring the force down to the desired range. The data above shows that both PFT and Tapping work in the desired force range, and that it is straight forward to image the top PDES material in PFT.What you call a disadvantage; I see as an advantage, in that with PFT, you can additionally simply increase the direct interaction force and now penetrate through the top layer, and image the previously hidden structure. If it were not possible to do the low force imaging, I would agree with your point, but since we can do both, plus get the direct measurement of height, along with all the material property channels, I see only benefit. We do agree that these are difficult things to do in Tapping Mode even with a stiff probe. That said, I love Tapping mode, and at Bruker we are committed to it and its continued development. With PeakForce Tapping, we are simply finding that in almost every case PFT provides better, easier, or more informative results.Best,Steve
Yes the images really are pretty impressive I have to say
Ethernet First Mile from the colocation providers Xifos.