A revolutionary imaging system could provide scientists with the ultimate microscope, one that can see inside very small objects with incredible clarity, but without relying on good quality lenses.

Diffraction was first exploited at the beginning of the twentieth century to size up atoms in crystals. Since then, X-ray and electron diffraction, and more recently neutron diffraction, have revealed the atom-by-atom structures of increasingly complex systems from simple
minerals to proteins, such as insulin, haemoglobin, and the genetic material DNA to clusters of proteins and the active sites of biological molecules as they interact with drugs.

Diffraction works because X-rays and electrons can act as waves and their wavelength is smaller than the spacing between the atoms of a solid. As these waves bounce off the atoms, they scatter and interfere with each other producing a pattern characteristic of the arrangement of the atoms. However, diffraction only works for crystals with their regular arrangement of atoms. To investigate more complicated atomic structures, scientists turned to electron or X-ray microscopes.

In conventional microscopy, a series of magnifying lenses focuses light coming through an object so that the microscopist gets a much closer view of the object than is possible with the naked eye. This approach is equally applicable to electrons or X-rays transmitted through the object. However, the quality of the lenses is critical to achieving a high-resolution image. Lenses for X-rays and electrons are extremely hard and costly to manufacture. Imperfections in the lens can ruin the image, making a typical electron or X-ray microscope picture about one hundred times more blurred than is theoretically possible for such a device.

In this project, the researchers plan to sidestep the intrinsic flaws in lenses by applying an entirely different approach to building the magnified image that will produce the best-ever pictures of individual atoms in any structure even if it is not crystalline.

Their approach uses a conventional X-ray or electron lens of relatively poor quality to form a patch of moderately-focused illumination. In fact, no lens is needed, just a moveable aperture near the object. Then, rather than "looking" at the object through a lens, they instead record the intensity of the diffraction pattern which emerges from the other side of the object on a good-quality high-resolution detector. They repeat these measurements for several positions of the illuminating beam. The result is a series of complicated diffraction patterns that in no way resembles the object itself, but a computer can then convert this digitised diffraction pattern using a calculation known as a "phase-retrieval" to extract the information. In this way, it should be possible to make much better images than with conventional lens-based systems.

The method could be implemented in existing electron and X-ray microscopes, greatly enhancing their imaging capability. It is possible that the research could also lead to a solid-state optical microscope, built into a single chip with no optical elements at all.

The new method will allow substantial advances in conventional optical imaging: by not requiring a short working distance (our method uses a low numerical aperture lens with a large working distance) and by being able to reconstruct all the scattered wave information computationally, the technique should compete favourably with the scanning optical microscope, but allow important benefits, such as imaging through thick enclosures (such as pressure cells), without the usual limitations of either shallow field of view or loss of resolution. Such methods could benefit many areas of research.

Practical techniques developed in X-ray and electron microscopy will be of direct benefit to solid-state physicists, materials scientists, biologists or nanotechnologists who need to know structural information at very high resolution, down to the atomic scale.

It is at the atomic scale where the greatest scientific benefits will accrue from this work. Low dose methods in electron and X-ray work could revolutionize structural determination of proteins which are difficult to crystallize: currently a major impediment in molecular biology, and hence the pharmaceutical industry. We anticipate that the costs of both electron and X-ray microscopes could be reduced by a substantial factor, without loss of performance, leading to their more widespread use in the analytical sciences. The ever decreasing size of semiconductor and nanoscale devices is now hampered by the extreme costs of high-performance microscopic techniques: our technology would benefit all such research laboratories.

Novel microscopies, using new forms of radiation, which are currently regarded as 'low resolution' methods (for example, low energy electron microscopy, XPS or acoustic imaging) would be improved by the work described, benefiting studies in new areas of structural analysis, testing, non-invasive imaging, forensic analysis and other areas of analytical imaging.

Innovations leading from discoveries made by this enabling imaging technology will potentially impact upon production methods in nanotechnology, materials processing and biotechnology. This will lead to broad benefits to society at large, perhaps leading to cheaper, more reliable consumer devices, and more effective medical treatments.