Analyzing the market for copper tools: From niche to mainstream?
05/01/2000
The front-end equipment market for copper processing of IC interconnects grew slightly more than 36% in 1999, according to our recent analysis. With 300mm on the back burner, we predict that copper technology will be the next market mover, perhaps growing more than 50% in 2000. Due to technological problems, copper equipment will probably remain a niche market for some time to come. It is likely to become mainstream within the next five years, however, moving into high gear with 0.15 and 0.13µm manufacturing.
The market
The copper equipment market, which includes copper electroplating, CMP, barrier/seed deposition, dielectric and metal etch, dielectric deposition, metrology, and RTP, will be worth slightly less than $3 billion in 2004, as shown in the table.
In 1999, Semitool was the market leader with a 40% share of units shipped, followed closely by Novellus with a 35% share. Applied Materials was next, while EEJA, Ebara, and CuTek (recently acquired by Steag) each shipped one unit. The installed base of electroplating tools showed Semitool with 55%, followed by Novellus with 30% through the end of 1999.
Trends
 |
The benefits of copper are well known, but chip manufacturers underestimated its potential problems when they first evaluated the technology. Thus, in early 1998, several copper-electroplating companies we contacted erroneously predicted that market growth for 1999 would be between 55% and 70%. Most companies had planned to transition to copper at the 0.15 or 0.13µm design nodes, but because of early developments in copper technology at IBM, Motorola, TSMC, and UMC, that transition was pushed up to 0.18µm, primarily for fear that the demand for high-speed ASICs would push ASIC suppliers to foundries that could offer the technology. Reports of low yields in Cu processing, however, forced several potential players to adopt a wait-and-see attitude, postponing its universal acceptance as a technology of choice.
Such pitfalls are rooted in the equipment industry's lack of a cohesive strategy and direction. A technology driver in the late 90s, copper was introduced with more hype than substance before customers were ready to accept it. Semi's (and Philips Analytical's) mid-1999 survey indicated that the "perception of 'readiness' was estimated to be lower by chip manufacturers than it was by equipment suppliers." Equipment suppliers, therefore, are spending vast amounts of R&D money on tools that the customer — the chip manufacturer — is not ready to buy.
Copper problems
The most significant problems encountered with the copper-dual-damascene process center around copper deposition and copper CMP. Equipment suppliers are working to enhance the quality of their tools, processes, and materials to address these problems.
Copper deposition. Semitool, in conjunction with Shipley, has developed a seed-layer-enhancement chemistry and process (patented by Semitool) that enables its LT copper-electroplating platform to correct deficiencies in the seed layer from PVD tools. Further, because seed-layer enhancement will overcome deficiencies, present-generation PVD equipment can be used on advanced generations of devices. The process can also influence CMP, since it provides a more consistent copper fill. The company's commanding market lead, in our analysis, is due to this technology, which is one or two generations ahead of its competitors.
CMP. This process is more complex for copper than for other metals because it is necessary to remove the tantalum or tantalum nitride barrier layers and copper uniformly without overpolishing any features. Copper's physical properties add to the polish difficulties: Unlike tungsten, it is a soft metal and subject to scratching and embedded particles during polishing. If a traditional one-step process is used to planarize copper, dishing (overpolishing) of the copper results as the pad reaches the much harder Ta or TaN barrier.
Several slurry suppliers have developed a two-step approach to copper CMP in which the first step uses slurry to planarize the surface quickly and remove copper uniformly down to the barrier layer. The second polishing step is performed with slightly modified slurry that removes the barrier layer without causing unacceptable dishing.
Other slurry solutions include EKC's MicroPlanar CMP 9000 series of copper slurries; Rodel's Copper-s1 Step 1 for fast bulk copper and Copper-s1 Step 2 for fast barrier removal, designed for use with the company's IC-1000 pads; Cabot's Epic 10-in. polishing pad for copper applications; and Strasbaugh's three-step copper CMP process in which the last step is an oxide buff to meet dishing and erosion requirements.
Copper contamination. This is a problem being explored by several equipment suppliers. During deposition, a film of copper on the wafer extends to the beveled edge and the wafer backside. Failure to remove all the copper can result in cross contamination in the fab.
Semitool's Capsule removes the PVD- and electrochemical-deposition backside, bevel, and edge copper contamination in situ, so wafers leave the tool clean. SEZ has developed a spin-etch cleaning technique to remove copper contamination from the front-side exclusion zone and beveled edge.
Copper film purity. One issue inherent in electrochemical deposition of copper is the purity of copper films. Customers want purer films and fewer defects, with the focus always on materials' cost/wafer. There is a price increase of 20-25% between 5N and 6N pure Cu; between commercial grade Cu and 5N, the price increases 40%.
Copper cathodes are also "tuned" for electroplating by adding phosphorus to control the microstructure. It also acts as a catalyst, preventing the formation of insoluble Cu2O instead of soluble CuO. Phosphorus concentration is in the range of 0.025-0.06%.
While there is a trend toward higher purity in the sputtering target market, the Cu anode purity issue stands out. Why deposit a seed layer of 6N pure copper when the interconnect is deposited from an anode with a 4N purity and the 6N target is priced 75% or more higher than the anode?
If the industry recognizes that low-purity copper does not affect a device deleteriously, it could mean a similar shift in thinking about other high-purity materials. There has never been a definitive study correlating the relationship between impurity levels at the ppb or ppt range, but engineers continue to press for higher-purity chemicals and materials with every successive generation.
Robert N. Castellano is president of The Information Network, a consulting and market research company addressing the semiconductor, computer, and telecommunications industries. For further information about the report 300mm/Copper/Low-k Convergence: Timing, Trends, Issues, Market Analysis, contact him at The Information Network, 8740 Lyon Valley Rd., New Tripoli, PA 18066; ph 610/285-4548, fax 610/285-4547, e-mail [email protected], www.theinformationnet.com.