Imaging with an AFM is SLOW!    When a high resolution image (say 512 x 512 pixels) needed over 8 minutes to complete, reading material was almost a requisite when conducting AFM imaging.   Not much you could do to get around that time needed to collect an image at a “typical” rate of 1 scan line/second or 1 Hz; if you had a really rough or tricky surface you would have to slow down to the interminable 0.5 Hz or less.  When AFM is invariably compared to other nanoscale microscopy based methods such as SEM and TEM, the AFM imaging speed is often cited as a serious drawback and even hindrance.  This criticism is a bit unfair because high resolution SEM imaging is also slow, but SEM does have the ability to screen at low res and much larger areas very quickly.

Fortunately there have been a lot of improvements in the imaging speeds of AFM, especially in the past decade.   There were a lot of technological hurdles to overcome to improve imaging speeds:

One of the biggest hurdles, not surprisingly, was the probe.   Remember that the probe is the heart of the AFM and will interact with your sample.  Increasing cantilever bandwidth was critical to the advent of high speed AFM, and this was basically accomplished with a robust, commercial-scale production of ultrasmall cantilevers.  Ultrasmall cantilevers increase the bandwidth by reducing the cantilever mass.  Let’s compare conventional AFM cantilevers with the ultrasmall ones designed for high speed application.   The specs of a standard “stiff” cantilever for topography imaging on polymers such as Olympus AC 160 probes are 160um x 40um x 4um (length x width x height).  Compare that to an Arrow UHF probe from Nanoworld that runs 35um x 42 um x 0.7um.   Images of a top view and the underside view of this probe are shown below – note the triangular geometry of the cantilever.  Also impressive are the resonance frequencies – the AC 160 probe has a resonance of around 300 kHz whereas the Arrow UHF has a resonance at about 2MHz!!

The Arrow is a high resonance frequency probe, courtesy of Nanoworld

Smaller cantilevers mean smaller optics for the optical beam deflection system!  Unfortunately reduced spot size also leads to reduced deflection sensitivity so you need a different laser that is optimized for this operation over use of conventional cantilevers.   So instead of a conventional 10 x 30um spot size as found, for example, on the Asylum Research Cypher, you have a specialized small spot laser module with a nominal 3 x 9um spot size instead.  Other technological hurdles included an AFM feedback loop with a z scanner capable of handling high imaging speeds and high speed electronics to go along with it.

So how fast is fast? 20x faster imaging speeds can now be achieved routinely with AFMs with high fast scanning options that can image at rates of upto 20Hz. Most research grade AFMs have a high speed scanning option including the Asylum Research Cypher, Bruker Dimension Fast Scan, JPK NanoWizard UltraSpeed and Park NX (this list is not exhaustive).

Fortunately the improvements in scanner and electronics have improved imaging rates on research-level AFMs so that user can routinely image with conventional cantilevers at double or triple the rates they used to, which has been a very useful improvement.  

So now that we have this exciting fast scan capability, and it’s been around for almost 10 years, can we say how useful it is??  It is clear that high speed imaging has been incredibly useful to study time-dependent processes such as kinetics and dynamics involving growth, melting, crystallization, and annealing of a wide variety of surfaces including crystals, polymers, and biological molecules.    Don’t forget that these processes can be studied under fluids too so the combination of fluids + temperature stage + high speed is powerful.   You can find some really pretty videos taking advantage of this high speed imaging capability in polymer and biological applications in the videos gallery of Professor Jamie Hobbs of University of Sheffield.  Since I cannot show videos on this blog, see below a series of 3 snapshots I took from one of his videos showing oriented polyethylene lamellae crystallizing at 128C.  For these specific research applications, the utility of high speed imaging is clear.

Snapshots of polymer crystalisation captured by high-speed AFM, courtesy of Jamie Hobbs

A few industrial labs that require high throughput take advantage of ultra high speed imaging.  But high speed imaging has not penetrated, for the most part, into the day-to-day AFM imaging.   Perhaps because of cost of the hardware and cantilevers, or getting involved with the dynamics of ultrasmall cantilevers, or because we are inured to imaging slowly, the superfast scanning at rates of 20Hz or so is not very common except for the niche applications above. 

By the way, faster is coming.  Publications of AFM imaging at rates of a few frames/second have been recorded by the laboratory of Mervyn Miles and commercial instruments with that capability are coming soon so stay tuned!

Dalia Yablon, Ph.D.

SurfaceChar LLC»

Imaging with an AFM is SLOW!    When a high resolution image (say 512 x 512 pixels) needed over 8 minutes to complete, reading material was almost a requisite when conducting AFM imaging.   Not much you could do to get around that time needed to collect an image at a “typical” rate of 1 scan line/second or 1 Hz; if you had a really rough or tricky surface you would have to slow down to the interminable 0.5 Hz or less.  When AFM is invariably compared to other nanoscale microscopy based methods such as SEM and TEM, the AFM imaging speed is often cited as a serious drawback and even hindrance.  This criticism is a bit unfair because high resolution SEM imaging is also slow, but SEM does have the ability to screen at low res and much larger areas very quickly.

Fortunately there have been a lot of improvements in the imaging speeds of AFM, especially in the past decade.   There were a lot of technological hurdles to overcome to improve imaging speeds:

One of the biggest hurdles, not surprisingly, was the probe.   Remember that the probe is the heart of the AFM and will interact with your sample.  Increasing cantilever bandwidth was critical to the advent of high speed AFM, and this was basically accomplished with a robust, commercial-scale production of ultrasmall cantilevers.  Ultrasmall cantilevers increase the bandwidth by reducing the cantilever mass.  Let’s compare conventional AFM cantilevers with the ultrasmall ones designed for high speed application.   The specs of a standard “stiff” cantilever for topography imaging on polymers such as Olympus AC 160 probes are 160um x 40um x 4um (length x width x height).  Compare that to an Arrow UHF probe from Nanoworld that runs 35um x 42 um x 0.7um.   Images of a top view and the underside view of this probe are shown below – note the triangular geometry of the cantilever.  Also impressive are the resonance frequencies – the AC 160 probe has a resonance of around 300 kHz whereas the Arrow UHF has a resonance at about 2MHz!!

The Arrow is a high resonance frequency probe, courtesy of Nanoworld

Smaller cantilevers mean smaller optics for the optical beam deflection system!  Unfortunately reduced spot size also leads to reduced deflection sensitivity so you need a different laser that is optimized for this operation over use of conventional cantilevers.   So instead of a conventional 10 x 30um spot size as found, for example, on the Asylum Research Cypher, you have a specialized small spot laser module with a nominal 3 x 9um spot size instead.  Other technological hurdles included an AFM feedback loop with a z scanner capable of handling high imaging speeds and high speed electronics to go along with it.

So how fast is fast? 20x faster imaging speeds can now be achieved routinely with AFMs with high fast scanning options that can image at rates of upto 20Hz. Most research grade AFMs have a high speed scanning option including the Asylum Research Cypher, Bruker Dimension Fast Scan, JPK NanoWizard UltraSpeed and Park NX (this list is not exhaustive).

Fortunately the improvements in scanner and electronics have improved imaging rates on research-level AFMs so that user can routinely image with conventional cantilevers at double or triple the rates they used to, which has been a very useful improvement.  

So now that we have this exciting fast scan capability, and it’s been around for almost 10 years, can we say how useful it is??  It is clear that high speed imaging has been incredibly useful to study time-dependent processes such as kinetics and dynamics involving growth, melting, crystallization, and annealing of a wide variety of surfaces including crystals, polymers, and biological molecules.    Don’t forget that these processes can be studied under fluids too so the combination of fluids + temperature stage + high speed is powerful.   You can find some really pretty videos taking advantage of this high speed imaging capability in polymer and biological applications in the videos gallery of Professor Jamie Hobbs of University of Sheffield.  Since I cannot show videos on this blog, see below a series of 3 snapshots I took from one of his videos showing oriented polyethylene lamellae crystallizing at 128C.  For these specific research applications, the utility of high speed imaging is clear.

Snapshots of polymer crystalisation captured by high-speed AFM, courtesy of Jamie Hobbs

A few industrial labs that require high throughput take advantage of ultra high speed imaging.  But high speed imaging has not penetrated, for the most part, into the day-to-day AFM imaging.   Perhaps because of cost of the hardware and cantilevers, or getting involved with the dynamics of ultrasmall cantilevers, or because we are inured to imaging slowly, the superfast scanning at rates of 20Hz or so is not very common except for the niche applications above. 

By the way, faster is coming.  Publications of AFM imaging at rates of a few frames/second have been recorded by the laboratory of Mervyn Miles and commercial instruments with that capability are coming soon so stay tuned!

Dalia Yablon, Ph.D.

SurfaceChar LLC «