Computational creativity and the future of AI

New process could revolutionize electron microscopy


March 20, 2012

A ptychographic reconstruction of gold particles showing the atomic fringes (Image: Univer...

A ptychographic reconstruction of gold particles showing the atomic fringes (Image: University of Sheffield)

Researchers at the University of Sheffield have created what sounds impossible - even nonsensical: an experimental electron microscope without lenses that not only works, but is orders of magnitude more powerful than current models. By means of a new form of mathematical analysis, scientists can take the meaningless patterns of dots and circles created by the lens-less microscope and create images that are of high resolution and contrast and, potentially, up to 100 times greater magnification.

In a recent paper published in Nature on March 6, 2012 under the daunting title of “Ptychographic electron microscopy using high-angle dark-field scattering for sub-nanometre resolution imaging," University of Sheffield scientists M.J. Humphry, B. Kraus, A.C. Hurst, A.M. Maiden and principal investigator John M. Rodenburg outlined their achievements in overcoming some of the limitations that have held back the potential of the electron microscope since its invention by Max Knoll and Ernst Ruska in 1931.

They demonstrated that the way to improve the electron microscope was by removing the thing that is at the very heart of the device - the lens. The difficulty in creating an electromagnetic lens of sufficient quality to allow the electron microscope to work close to its theoretical limits have kept it operating at magnifications 100 times less than what it could be. They found that the way to improve the electron microscope was to eliminate the lens and replace it with a virtual lens, created by applying a new form of mathematical algorithm to diffraction patterns.

Even though electron microscopes have been around almost as long as motion pictures have had sound, they remain mysterious things. Large, complicated and expensive, they are inhabitants of the research laboratory, medical facilities and high-tech industrial firms. To the public, they are the source of alarmingly vivid images of tiny insects turned into monsters from a B-movie or micro-circuits transformed into modern architecture. Yet despite looking like a prop from a science fiction film, the principle behind them is identical to that of the common, garden-variety optical microscope of the kind found in school laboratories and toy shops.

The type of optical microscope most commonly used in schools is called a transmission microscope. In this, light is produced at the bottom of the device by a lamp or mirror. The light shines upward through a glass slide where the specimen under examination is mounted. The light passes through the specimen, then through a series of lenses that magnify the image. These lenses may look very complicated, but the principle is that of an ordinary magnifying glass - refracting light waves to produce a larger image. Indeed, the simplest microscopes only use one lens. The compound lenses found in more complex scopes are there simply to increase the magnification within a reasonably small instrument, and to correct any optical distortions so that the final image is clear when it reaches the eyepiece or imaging camera.

An electron microscope works exactly the same way, only it uses a beam of electrons instead of light. This is because the amount of magnification an optical microscope can achieve is limited by the wavelength of light. Visible light waves are relatively large, so there’s an inherent limit to how much they can magnify something - a bit like a low-resolution digital camera. Try using the digital zoom on a cheap phone camera and you only get a bigger, blurrier image. Optical microscopes can magnify objects up to 2,000 times. Above that, the light waves start interfering with one another and the images become blurry and surrounded with bright rings formed by the interfering waves.

There are a number of ways of getting around this limitation. The simplest of these is to replace visible light with shorter wavelengths, such as in ultraviolet microscopes. The shorter the wavelength of the light waves, the smaller the objects the microscope can see. The electron microscope takes this idea much further.

Electrons are solid particles, but they are so tiny that, like photons in light, they act like waves. However, unlike visible light, electrons can form waves of much smaller length, so they can achieve much higher levels of magnification - up to 10 million times as opposed to an optical microscope’s 2,000. The only snag is that glass lenses are useless when it comes to focusing electrons, so a different kind of lens is needed. It’s one that isn’t made of any sort of material, but rather a lens of pure energy. For the electron microscope, the lenses are made of electrostatic or, more commonly, electromagnetic fields. These fields are generated by the microscope and manipulated to form lenses, which are arranged exactly as they are in a compound microscope.

There are several kinds of electron microscopes, but for our purposes, we’ll look at the transmission electron microscope (TEM). The TEM is a bit like the school microscope, only built upside down. Instead of the light shining up from a lamp at the bottom, there is an electron gun installed at the top of the scope. This gun fires out high-voltage electrons, which are accelerated by an anode and pass through an aperture to create a beam. This beam passes through a pair of condenser lenses, passes through the specimen under examination and through an objective lens to focus the image, which falls on a fluorescent screen, photographic plate or digital array.

Despite this similarity to an optical microscope, the TEM looks rather intimidating. Part of the reason is all the controls needed to control the high-voltage equipment used to generate the beam safely. The other is that the TEM needs to work in a vacuum, which means that it must incorporate a vacuum chamber, the pumps used to remove the air, and airlocks for placing specimens inside the machine. But underneath, it’s just a matter of lenses.

The problem with electron microscopes in general is that they never operate anywhere near their theoretical limit. Though they are capable of incredible powers of magnification, in actual operation the microscopes can be anywhere from 25 to 100 times less powerful than they could be. This is because of the great difficulty of building lenses that can exploit the full potential of the electron beam. The slightest imperfection in the construction of a field coil, the smallest fluctuation in the power going to it, and the electron microscope gets a case of electromagnetic astigmatism. There are simply too many variables that need to be under too tight a control to make these lenses perfect.

One way to get around this is the brute force method - that is, boost the power of the electron beam. This allows some improvement, but pumping more energy into a specimen tends to cook it like a microwave burrito, which is not good.

The alternative is the one that Rodenburg and his team at the University of Sheffield opted for - get rid of the lenses altogether. This may sound counter-intuitive or even like nonsense, but it does work because in taking out the lenses the Sheffield team has replaced them with a virtual lens that isn’t made of glass or electromagnetic fields, but of mathematics.

The secret behind this is diffraction, which, ironically, is what causes those bright rings of light around objects that help limit the power of optical microscopes. If it’s been a while since physics class, diffraction is what happens when any wave encounters an obstacle such a stick in the water, the pinhole in a pinhole camera or other waves. As the waves try to pass around the object they start to crowd and interfere with one another. This interference or diffraction follows set rules and creates patterns that carry a great deal of information that scientists and engineers can use.

One of the simplest examples of this is x-ray crystallography. Crystals have a very regular structure and when an x-ray beam is fired at the surface of one, the x-rays are scattered and diffracted and create patterns on photographic plates. From these patterns, scientists can work backwards and calculate what kind of crystal structure would produce it. It’s this sort of data recovery that the University of Sheffield uses to create its virtual lens.

Imagine taking a standard biology microscope and removing the lenses, but leaving a aperture to create a light beam. You pass this beam through your specimen slide and let it fall on a high-resolution detector. You won’t get a proper image, but you will get a diffraction pattern created by the beam passing through and being diffracted by the specimen. This pattern carries information about the specimen, but not enough to be useful. From the pattern, it’s possible to measure the intensity of the waves that created it, but the timing of when the peaks and troughs of the wave reached the pattern (their relative phase) is lost.

The clever bit is to move the aperture of the microscope along the specimen and take several shots that build up a much more complicated pattern. This combined pattern can then be subjected to a mathematical algorithm called ptychography that can perform what is called “phase retrieval," which recreates the lost phase data from the pattern. The end result is an image that is more accurate and of higher resolution than that produced by a real lens.

This technique is not itself new. It was first applied to optical microscopes in 2006 with considerable success. Sheffield’s achievement is to take this a step further by applying it to electron microscopy. This not only allows them to overcome the limitations of electromagnetic lens and gain resolutions down to a potential of 1/100th of an atomic diameter and to provide much more detailed images, but it provides a truly virtual lens that it not only more powerful, but also much more flexible. By manipulating the algorithms used in ptychography, the calculations can change the depth of field and focus. This should allow scientists to zoom in on any desired feature of the specimen and study it to the best perspective, much as a photographer can change settings on a zoom or macro lens for the best shots.

The practical benefits of the lens-less electron microscope are considerable. The advantages of increased magnification being available to scientists and engineers are obvious, but by achieving this without increasing beam power means that, for example, delicate protein molecules can now be studied without destroying them in the process. The higher resolution also pushes back the boundaries of micro-circuitry.

This just scratches the surface of the potential of such virtual lenses. The lens-less optical microscope already promises a future of microscopes on a chip. Time will tell what the lens-less electron microscope will lead to.

Source: Nature

About the Author
David Szondy David Szondy is a freelance writer based in Monroe, Washington. An award-winning playwright, he has contributed to Charged and iQ magazine and is the author of the website Tales of Future Past.   All articles by David Szondy

This is science..

Daniel Brown
20th March, 2012 @ 03:14 pm PDT

That is a well written article.

Long enough to cover the subject, detailed enough to illustrate the principles and short enough to not be an encyclopedia on the minteau and intricacies.

I wonder when the next ramp up in smallerness than evera will be? Like what is way smaller than electrons and they are reasonably easy to create and use?

Mr Stiffy
20th March, 2012 @ 04:57 pm PDT


Great article!! I'm an engineering PhD student, so I'm familiar with SEM and TEM. This are great news, with this method the application field of this tools is going to be even larger.


20th March, 2012 @ 08:20 pm PDT

taking this farther, but from a non-tech, real-time viewing would be a plus and after that shooting at a particle to get a result would be a plus. sort of like colliding atoms at high velocity, but instead to get a particular result.

21st March, 2012 @ 06:24 am PDT

It would be nice to have a scale for the photo.

It's even nicer to see user comments that highlight the flaws in such systems.

Often one is greatly impressed with innovations, then Gizmag readers with their assorted experiences and background shoot them down to size, which is illuminating and educational. Having said that, this sounds like a brilliant and revolutionary development. But it reminds me of disappointing blurred 20 micron TEM photos I used to get of polymer chains (for Ph.D. work) until I realised the blurring was due to actual thermal vibration of the molecule. I wonder if this new system will require cryogenic cooling of the samples every time.

24th March, 2012 @ 10:47 pm PDT
Post a Comment

Login with your gizmag account:

Or Login with Facebook:

Related Articles
Looking for something? Search our 31,306 articles
Recent popular articles in Science
Product Comparisons