High resolution extended image near field optics

1. Introduction and Conclusions

[this page | pdf | references | back links]

Copyright (c) Malcolm Kemp 2010


Return to Abstract and Contents

Next page


How practical it is to achieve arbitrarily high resolution with a microscope, telescope or photolithographic device?


In days gone by, this topic would have had a clear answer. The wave-like nature of light places a limit on the effective resolving power of conventionally formulated (i.e. ‘far field’) devices. This limit is given by the well-known Rayleigh’s equation, which states that the minimum distance able to be resolved by such a microscope or a photolithographic device, , is set by:



In this formula,  is the wavelength of the light used,  is the numerical aperture of the device and  is a constant that depends on the image formation technique used and (for microscopy or photolithography) the specific resist being used in the device. A similar equation, involving the aperture size and wavelength limits the angular resolving power of a (conventional) telescope.


However, given the huge technological and economic benefits accruing from squeezing ever finer circuitry onto microchips, cunning researchers have squeezed down the value of  using a variety of techniques. Some have involved improving  using ‘unconventional’ designs, some have involved shortening the wavelength of light being used, and some have involved lenses with enhanced numerical apertures, see e.g. Ito and Okazaki (2000) for further details.


Perhaps the most extreme example of an ‘unconventional’ design is the scanning near field optical microscope (‘SNOM’). This works in a manner similar to a scanning tunnelling electron microscope, in that the microscope is placed a fraction of a wavelength away from the object being imaged. The Rayleigh equation no then longer applies and arbitrarily accurate resolution is in theory possible (even to the level of individual molecules or, with scanning tunnelling electron microscopes, individual atoms).


SNOM’s operate by shining light through a very small hole which is placed very close to the object being imaged or projected onto. In principle, SNOM’s have arbitrarily high resolving power, since in theory their accuracy is limited only by the size of the hole through which the light is shone. In practice, however, their resolution is limited because of the finite skin depth of real metals, or in other words because idealised metals with infinitely high (complex) refractive indices do not actually exist. This places a lower limit on the effective hole size that it is possible to achieve in practice (although it is possible circumvent this limit to some extent by arranging for the light source to be a single molecule).


It is possible to use the same approach in reverse for lithography, although Ito and Okazaki (2000) express the view that “although such approaches are useful for producing individual nano-structures for the investigation of nanometre-scale devices, the throughput is likely to always remain impracticably low for commercial application”.


What is less clear is whether near field optical lithography (or microscopy) necessarily has to operate a point at a time, and therefore whether the assumed commercial barrier indicated by Ito and Okazaki (2000) is correct. We show that this assumption is probably false and that a resolving power similar to that achievable by existing SNOM designs should in principle be achievable by a near field device that creates an extended spatial image ‘all at once’. We shall also describe how in principle this approach might be used to create a telescope that also circumvents the Rayleigh resolution limit, despite how far away any practical object being viewed will be from the viewing telescope.


Whether these ideas will ever prove practical and/or commercial is more difficult to say – they were originally set out in a patent application in 2001, see Kemp (2001) and have not developed further since then. Other arguably more commercial techniques have since been developed that also ‘stretch the envelope’ of what is possible with high resolution optical lithography. However, continued improvement in resolving power is one of the key drivers needed to maintain Moore’s law and the microchip revolution we are all benefiting from. Perhaps, one day, the ideas set out in these pages will therefore prove useful and fruitful, if other methods of stretching the envelope run out of steam.


Contents | Prev | Next

Desktop view | Switch to Mobile