Scanning probe microscopes are a very sensitive class of microscope and work according to the following principle - a very sensitive and very sharp tip (the probe) is finely scanned over the surface of an object, or - actually - the object finely scanned under the probing tip (remember the old vinyl-record players). The interaction between the probe and the object's surface is recorded by very sensitive detection devices. The strength of the interaction is then plotted against the scan co-ordinates thus resulting in an interaction map, which corresponds to the surface's topography. Three factors determine the resolution of such an image:

• The scanning accuracy
• The tip's sharpness
• The interaction's strength

The scanning method is what all techniques roughly have in common. It is not the bottleneck factor for resolution. The scanning can be performed with an accuracy of some hundred picometres (that's about the width of one atom) thanks to the development of very precise positioning motors (piezo-motors) and high-precision positioning accuracy (using capacitance).

The tip's sharpness will tell how rounded-off the topographical images will look like. If a very sharp and long 'needle' is used then the cliffs and valleys on the object's surface will look sharp. If the tip is round and short then the images will look more blurry. The tips can be manufactured in various ways depending on specific requirements, ie, must the tip be soft, hard, sharp, coated, conducting, insulating, etc... Methods to manufacture tips range from electron beam deposition or sputtering to etching and lithography (as in microchip manufacturing).

The kind of probed interaction is what will distinguish the different scanning probe techniques. In all cases the interaction to be measured must be chosen so that the faint differences in intensity can be detected. The intensity of the measured interaction must significantly outweigh the background noise to yield satisfactory images at the highest possible resolution (ie in order to 'see' every atom of the surface).

Scanning Tunnelling Microscopy (STM)

The 'scanning tunnelling microscopy' or STM was the first scanning probe technique to be developed. It was developed by a group of scientists led by Gerd Binnig and Heinrich Rohrer in 1981 at the IBM labs in Zurich, Switzerland (both won Nobel prizes in 1986 along with Ernst Ruska, inventor of the electron microscope). This technique uses a very improbable quantum effect called 'tunnelling' as a measure for the distance between the tip and the surface of the object. Quantum effects take place on an atomic scale of length, and decrease dramatically with distance (distances in this scale are comparable to atomic radii, or tenths of nanometres). For this reason, when a tip is some atom's diameters away from an object, the electrons of the last atom of the tip sense the electrons from the outer atoms of the object's surface, even allowing the flow of electrons. Since this flow is classically not allowed, the electrons seem to 'tunnel' through space from the surface to the tip. This flow can be measured as a tunnel current (some billionths of Amperes). So, 'all' Binnig and Rohrer did, was to bring a metal tip close enough to a metal surface, scan them over the surface and measure the current at every spot, and then look at the image.

Why Metal Surfaces?

For the tunnelling quantum effect to take place in a measurable range, the electrons of the object's surface must be 'prone' to tunnelling. That is, they must be loosely bound in the first place. This is the case with metals. Secondly, the current between the tip and the surface must be measured, so - roughly speaking - there's a wire attached to the surface. This would not make much sense if the surface didn't conduct electricity. Unfortunately, this is also the main limitation of an otherwise amazingly easy technique (see Make yourself an STM). However, there has been some effort to push the limits further, semi-conductors and some metal oxides (which are not conducting in a strict sense) can be observed with an STM. Obviously, the 'tip' must also be made out of a conducting material.

STM in Practice

An STM is a little bigger than a king-size flip-top cigarette box with wires going to a computer. This box is internally vibration-damped and accommodates the piezo motorised sample-holder and a piezo motorised tip-holder. The sample-holder and the tip-holder are connected to the sensitive current-measuring device, which will output current values to the computer. Any metal can be used to act as a tip (etched iron wire tips, or copper tacks for example), but the best results are obtained with special tips made out of alloys containing iridium and platinum. All the piezo motors are also connected to and seated by the computer. With all this information (say, the position of the motors and the currents) the computer can reconstruct a topographic image of the sample's surface.

To operate an STM, normally, there is a so-called 'bias' voltage of around 1 V applied between the tip and the sample, so that the electrons 'know' from where to where they are supposed to tunnel. Applying higher 'bias' voltages of some 10 V can force entire atoms to jump from the surface to the tip. By reversing the potential, ie, applying -10 V the atoms jump the other way round from the tip to the surface. Using this neat little effect, one can place atoms on a metal surface at any desired position. This has been done to fabricate the famous IBM logo made out of 35 Xenon atoms (see it at IBM STM gallery). At least by then, the scientific community went ape about the STM.

STMs operate in two modes - constant height or constant current. In the first case, the tip-piezo-motor is held at a given height and the topography¹ is given by the different measured tunnelling currents. In the second case the tip's piezo motor is adjusted by a feedback mechanism so that the tunnelling current remains constant at a given value (it moves up and down as single atoms, which look like close-packed marbles, pass underneath the tip). In this way the topology is given by the tip's distance to the object's surface.

Force Microscopy (AFM)

Binnig proved to be exceptionally creative. It was also him, along with his other IBM chaps, Christoph Gerber and Calvin Quate, who invented the 'scanning force microscope' in 1986. These microscopes are more commonly called 'atomic force microscope' or AFM, because it can also resolve single atoms on a surface, and because the abbreviation is easy to pronounce. AFMs are a sort of an extension of STMs to non-conducting materials, and work in a very similar way. The difference is that in the case of an AFM the mechanic deformation of a very sensitive cantilever - on which the tip is located - is measured² instead of the tunnelling currents. Note the analogy between this method and the good old vinyl record-players. The tips must not touch the surface. Like with the STM, when the tip comes very close to the surface it starts sensing the outer atoms of the surface (the close-packed marbles). At this point the tip forces the cantilever to start bending very little, and the major task is to detect this small mechanic bending.

Cantilevers and Tips

One big problem is the manufacturing of the tips and cantilevers: At first the same etched wires of the STM system had been used along with the same piezo-motor and capacitance distance measuring, but they did not achieve good reproducibility of experiments - because it's one thing to have uniform conductivity and another thing to have uniform mechanical properties. Good uniform tips are therefore the key to this technique. Modern tips are manufactured by the same lithographic process used to make microchips. Common tips look like pyramids which are couple of microns long (about the same at the base). At the tip, these pyramids are about 10nm broad. The tip (pyramid) itself is situated at the loose end of a cantilever, which is some tenths of a milimetre long. The tiny deformation of this cantilever is the final measured property.

There are many ways to sense the bending of such a cantilever. The most common method is to aim a sharp laser beam at the tip of the cantilever and then measure its deflection, a method called 'beam-bounce' introduced by G Mayer and N Amer in 1990. The deformation of the cantilever acts like the tilting of a mirror, a laserbeam reflected on the untilted mirror will arrive at a different angle on the detection plane as a beam reflected by a tilted mirror. From this angle the bending of the cantilever, and thus the tip height, can be calculated.

AFM in practice

An AFM looks like a big STM, because now one needs lasers and bigger detection gizmos. Commonly an AFM is mounted on an optical microscope to aid with the laser-beam aiming. The rest is easy, the wires coming from the piezo-motors and from the detector go into a computer, where the images are reconstructed - same procedure as with the STM.

There are four modes in which one can run an AFM experiment. The modes are called contact, non-contact, tapping- and force spectroscopy modes. The first three methods are used to acquire images, the fourth method is used to probe the surface at a certain point. In the first two imaging modes, it is also possible to using 'constant force' or 'constant height' variations, both work like the STM variations described above. In contact mode, the tip touches the surface and is deflected as the atoms and other features of the surface pass underneath it (like a vinyl record-player). This mode is also called repulsive mode. The other imaging method, the non-contact mode, is also called attractive mode because the deflection of the AFM tip is caused by the electrostatic attraction of the cantilever tip and the weak transient charges on the surface of the sample (AKA van-der-Waals forces). In tapping mode, the tip oscillates vertically and 'taps' the surface. This reduces the force exerted on the surface so one can image nasty glorp like polymers, cells, etc. In force spectroscopy, the tip touches the surface and is lifted. As the tip interacts with the surface when it's lifted, it is slightly bent downwards. This poses a very sensitive method to sense absurdly weak forces. In this mode one can evaluate how sticky a surface is. Furthermore, by preparing the tip and the surface, this mode can be used to measure the rigidity of single molecules, the pulling force of protein motors or the stability of protein conformations (folding).

More Tips

Tips of AFMs do wear out, because some touching of the surface is inevitable. So with time, tips become rounded off (decreasing resolution). For this reason, a good tip-arsenal is of some importance. AFM tips are not as easy to manufacture (lithography) as STM tips (dunking wires into acid). Recently, a new tip concept has been presented. Here, carbon nanotubes (which are - in a first approximation - nanometre-sized carbon needles, which are very rigid and very flexible) are attached to worn-off AFM tips and then function as the new tips. Since they are very sharp and rigid the resolution is significantly improved. However, at the time of writing, these tips are still at an experimental stage.

NSOM - Another Scanning Probe Technique

For the sake of completeness this 'alternative' scanning probe method will be briefly described. The scanning and the overall working mechanism is the same: A sharp pointy something and a finely scanned object underneath it, in the end some interactions are measured and a map is visualised.

The near-field scanning optical microscopy (NSOM) works like this: A beam of light cannot be focussed onto an area which is smaller than half its wavelength. At least not with ordinary optical elements. Special glass fibres can be extruded so that its width becomes as wide as 10nm. In this way (laser) light can be forced to focus on such a small spot. However, some tenths of nanometres away from the glass fibre the light will dissipate to its common width (hence the name 'near-field'). This glass fibre is used as a tip, the light in the tightened focus (right after the tip) is used to interact with the surface of an object (scattering, absorbtion or fluorescence) which is being scanned underneath it. On the other side of the object, the interaction is detected by a conventional light microscope, and the image reconstructed with a computer. NSOM has the usual disadvantage of a scanning probe method, namely the restriction to a surface. Furthermore it is a very complicated method of very limited use. However, as it is not restricted to solid surfaces (as the STM and the AFM are), some insight into liquid stuff can be gained by using this method.