Even with Intel signaling its intent to bypass 157 nm, calcium fluoride will likely remain a critical optical material.
Konrad Knapp, Schott Lithotec AG
When Intel Corp. announced its intention to bypass 157-nm lithography last spring, it raised more questions than it answered for many in the lithography industry. However, it is important to recognize that this decision was most likely based on an assumption that key elements of the technology — such as calcium fluoride for optics, the pellicle and, to some extent, the photoresist — would not be available in the required quantities when needed.
In the past, the industry’s road map decisions have almost always been based on consensus. In fact, the few exceptions, such as IBM’s strong support of x-ray lithographic technology, never resulted in the expected benefits for their supporters. From a technological standpoint, there is no doubt that Intel could work with 193-nm technology and high-numerical-aperture (0.93) tools for the 2007 node (45-nm integrated circuit [IC] processes).
However, this would hardly represent a solution for the many chipmakers that specialize in DRAM because they will not be able to work with 193/0.93-NA technology to manufacture features with the required half pitch of 45 nm. If they are not allowed access to 157-nm technology, they either will have to extend 193-nm technology even further than Intel has or will be forced to switch to another technology that is unlikely to be available on time.
It should, therefore, be no surprise that Intel’s decision has not met with the complete support of the IC community. Because extreme-UV lithography is not expected to be a timely substitute for 157-nm technology, additional emphasis will be placed on extending 193-nm technology even further, by utilizing immersion techniques and optical lenses with a numerical aperture larger than 1.1. The 193-nm immersion represents a technological path that is expected to work for the 45-nm node.
Most recently, all exposure toolmakers have focused their developmental efforts on 193-nm immersion. Nikon entered into cooperative work on liquid immersion technology with Tokyo Electron Corp. Canon Inc. presented 193-nm immersion as a very attractive approach, and ASML Holding NV in Veldhoven, the Netherlands, updated its road map, considering immersion tools at both 193 and 157 nm. However, many question whether this can happen on time and whether the costs of commercializing very high numerical aperture lenses will be lower than the costs of continuing to develop 157-nm technology.
In Jena, Germany, Schott Lithotec produces calcium-fluoride crystals, with optical material available in a disk diameter of up to 350 mm and a thickness of more than 150 mm.
Clearly, when Schott Lithotec AG started developing manufacturing technology for the utilization of calcium fluoride at the 70/65-nm node back in 1997, its properties were far from meeting lithography’s high standards. Initially, the progress in material development could not keep up with the rapid pace of the industry because of the difficulty involved in working with the material. However, the improvements that have been made with calcium fluoride allow it to be used down to roughly the 55-nm node requirements. Furthermore, the latest developments will fully support its use at the 45-nm node and below.
Overcoming birefringence
The discovery of calcium fluoride’s anisotropic behavior at very short wavelengths, known as intrinsic birefringence, was clearly one reason for the delay in the material’s development. This forced optical designers to work intensively on new solutions to compensate for the loss in imaging performance of lenses made of birefringent material. It persuaded Schott Lithotec to engage in developmental efforts on lens blanks with a special (100) crystalline orientation.
Results of this successful initiative were presented to the lithography community at the 157-nm symposium in August in Yokohama, Japan. They suggest that the industry can sequentially take steps toward introducing lithography tools that will work at the 55- and 45-nm nodes. Obviously, this assumes that the other key elements of technology, such as the pellicle and the photoresist, also will become available on time.
It also should be noted that Intel’s decision to remove 157-nm technology from its road map has not met with many followers. In fact, the technologies under consideration for the 45-nm node and beyond — notably, 193-nm immersion and 157-nm and EUV technologies — compete with each other. Because all of these still require technological improvements and would thus incur considerably greater costs of ownership, the lithography community is hesitant to make a final decision. Apparently, exposure toolmakers are unable to receive firm commitments on either 193-nm immersion tools or 157-nm “dry” tools.
Over the next few months, 157-nm tools such as the Micrascan VII and the AT1600 step and scan tools from ASML will become operational. Schott Lithotec is certain that they will provide convincing evidence of the viability of this technology. This should help the industry to make clear decisions on the strategic value of higher-quality calcium-fluoride material.
Only then will people know whether it can provide enough momentum to the industry to make 157-nm lithography the technology of choice for the commercial and technological challenges to the IC industry at the 45-nm node and below.
Meet the author
Konrad Knapp is vice president of corporate business development at Schott Lithotec AG in Jena, Germany, where he is responsible for next-generation lithography.