 |
 |
17 Oct 2003
We recently reported Intel's decision to skip 157 nm lithography and focus on enhanced 193 nm and extreme ultraviolet (EUV) technology instead. In this article, the leading maker of 157 nm optical materials, Schott Lithotec, responds to accusations that calcium fluoride lenses will not be ready for chip production using the 45 nm node in 2007.
Following a series of improvements in calcium fluoride lens manufacture over the last few months, we believe that the industry will be able to use 157 nm lithography tools for the 55 and 45 nm nodes. However, it is important that the other key technology elements, such as the pellicule and photoresist, are made available in time.
Intel’s decision has narrowed the potential use of 157 nm technology and it seems that the community is unable to decide on a common technology to reach the 45 nm node and beyond. This is largely due to concerns over the high cost of ownership which is a common issue for each of the lithography choices – 157 nm, 193 nm immersion or EUV.
It is important to note that in the past, the semiconductor industry’s roadmap decisions have always been based on consensus. The few exceptions – such as IBM’s strong support of X-ray lithography – have not brought the expected benefits for the stand-alone supporters.
There is no doubt that Intel will now work with 193 nm technology and high (0.93) numerical aperture (NA) tools for the 2007 (45 nm) node. However, this will not be a solution for other chip makers who manufacture dynamic random access memory (DRAM). For these memory makers, it is unlikely that 193/0.93 NA technology will be able to manufacture features with the half-pitch of 45 nm that they need.
If 157 nm technology is dropped, these firms will either have to extend 193 nm technology further than Intel, or switch to another technology, which is even less likely to be ready in time. As a result, Intel’s decision has not received widespread support from the chip manufacturing community.
Race against the clock
Although 193 nm immersion lithography ought to work for the 45 nm node, there are some concerns. These are chiefly whether it can be developed in time, and whether the costs of commercializing the very high NA lenses required will be lower than the cost of continuing 157 nm development.
Certainly, calcium fluoride was a material with properties some way from lithography requirements when Schott Lithotec started to develop it for use with the 70–65 nm node in 1997. It’s also true that its development – with a low-productivity process – initially did not keep up with the pace of the industry, which had underestimated the technical complexity involved.
However, recent improvements now allow optical calcium fluoride to fulfil the requirements of the 55 nm node. Further developments which are currently under way will fully support its implementation at the 45 nm node and below.
One of the reasons for the delay in calcium fluoride development was the discovery of its anisotropic behaviour at very short wavelengths, which became known as intrinsic birefringence. This physical effect forced optical designers to work intensively on design solutions that would compensate for the resulting loss in imaging performance of lenses built with birefringent material.
In particular, it forced Schott Lithotec to pursue a critical development effort for lens blanks with a special (100) crystalline orientation. The results of this innovation were recently presented to the lithography community at the 157 nm symposium held in August in Yokohama, Japan.
In the next couple of months 157 nm tools such as ASML’s Micrascan VII and the AT1600 will become operational. These will provide evidence of the viability of 157 nm technology.
They will also allow manufacturers to decide whether the improvements in calcium fluoride can provide enough momentum in the industry to make 157 nm lithography the technology of choice for the 45 nm node and below.
 |  |  |  |  |  |  |
| © 2024 SPIE Europe |
|
