For the past 50 years, electron optics engineers have sought to improve the precision of electron microscopes by counteracting the image blurring effects of lens imperfections, or ‘aberrations’. To fix the problem, IBM and Nion researchers have created the world’s highest resolution electron microscope that uses innovative technology to obtain images of individual atoms inside a material at a resolution never achieved before. The breakthrough could improve computer chip materials and manufacturing by allowing much more detailed, atomic resolution, structural examinations of semiconductors to be carried out.
Key Elements of the New High Resolution Electron Microscope
As reported in the 8 August issue of the journal Nature, the new technique significantly extends the capabilities of the electron microscope. The new system combines seven major optical elements, adding three octupole and four quadruple lenses to an electron microscope. ‘This system achieves 20 times the wavelength of the 120kV electrons by using about 28 multipole electron optical elements added to the relatively simple, four lens column of a dedicated scanning transmission electron microscope,’ says Dr Philip Batson, the lead scientist on the project at IBM Research. ‘Computer measurement of optical aberrations and control of the many lenses in response are both key to getting the very complex system to work.’
How the New High Resolution Electron Microscope Works
The measurement of optical performance is done using a TV camera that views a ‘shadow map’ of randomly structured test samples. The shadow map is then analysed by a computer to produce corrections for the 28 or so lens coils, thereby optimising the system’s performance for a very small electron beam size. After this correction, the microscope can produce an electron beam that is smaller than one angstrom (1Å), or 75 thousandths of a nanometer. ‘For my field of semiconductor research,’ says Batson, ‘it is very significant that we can now image with 1Å resolution or better using electrons that have an energy below the knock-on displacement threshold of about 160kV for silicon.’
What Can be Imaged Using the New High Resolution Electron Microscope
As the dimensions of computer chips shrink, this new correction technique, co-developed by Niklas Dellby and Ondrej Krivanek from Nion Co, Kirkland, Washington, USA, will enable scientists to see, for the first time, a complete picture of the atomic structure of important defects in the materials used in these chips. Watching how atoms assemble, move around and interact with other atoms in silicon, for example, could help scientists towards a better understanding of how to control environmental conditions so components of future computer chips could assemble themselves. ‘The field of aberration correction will now move rapidly towards ease of use and correction of higher order aberrations,’ believes Batson. ‘We will see optical columns designed from the beginning to use aberration correction and computer control.’
Has the New Technology Overcome Accessibility Problems?
In the past, high-resolution systems were only available in a few locations because of their complexity and cost, putting them beyond the use of many individual investigators. While the new technology integrates a complex optical system, its computer measurement and control allows this complexity to be handled more easily. Batson is excited by the development and sees it as a move towards smaller, possibly cheaper systems.
Future for the New High Resolution Electron Microscope
Batson’s work is currently focussing on high-resolution electron energy loss spectroscopy, having integrated an electron monochromator and spectrometer into the system. This gives a sub-100meV energy resolution, which Batson is using to investigate the electronic structure of defects and interfaces in semiconductors. ‘The rather large question right now,’ he says, ‘is how well this activity can be integrated with the atomic resolution imaging to make a direct connection between electronic and atomic structures for the future design of novel semiconductor devices.’