Technology News
01/01/2000
Cleanroom airflow swirling in controversy
Recent studies by Maribel Vazquez of MIT's Department of Mechanical Engineering shed new light on the effects of airflow in a cleanroom. Vazquez challenged current approaches to cleanroom airflow practices by demonstrating the pitfalls of: assuming laminar flow, using high air velocities; and evaluating cleanliness by counting particles in the air.
|
Her improved approach to cleanroom design includes turbulent flow analysis, lower air velocity than the current convention, and measurement of cleanroom effectiveness by particle deposition.
These results of experiments and simulations were presented during the session "A Study on Altering Air Velocities in Operational Cleanrooms" at the recent CleanRooms West Show.
Vazquez created a three-dimensional array of 120 hot-wire anemometers in a cleanroom and measured the air velocity at a 100kHz sampling rate. The large quantity of data revealed the turbulent nature of the airflow. The "turbulence intensity," defined as the deviation from the mean velocity expressed as a percentage of the mean, was in the range of 2-8%, depending on the height and the distance from the walls. Anything greater than 2% is considered to be highly turbulent, so the flow was clearly out of the laminar regime.
The data also showed that the turbulence intensity increases with higher air velocities near the center of the room, making the laminar flow assumption most erroneous in the most interesting area of the room. The turbulent flow would trap airborne particles in eddies and then send them onto products when operators move through them.
The second experiment measured particle deposition rates on test wafers distributed around the cleanroom. The particles were half-micron diameter, NIST-certified polystyrene spheres, and a humidifier was used to inject them into the room. The table shows that, although the average deposition rate was higher at low air velocities, the low airflow decreased the deposition rate in the middle of the room, where the wafers would be processed. A lower air velocity not only keeps the critical area of the cleanroom cleaner, but it also reduces the cost of blowing air through the facility.
The outcome of this type of analysis would vary with the design of the cleanroom, but these results show that cleanroom airflow deserves more detailed study than the conventional wisdom dictates. J.D.
Wacker Siltronic introduces fLASH wafer
Wacker Siltronic AG, Germany, recently introduced a silicon wafer used in advanced microelectronic applications. The fLASH wafer, as it has been named, consists of a refined surface layer on a precipitation-optimized substrate.
"The first step in the production of the new flash was to produce a cost-effective substrate which precipitates in low thermal budget processes without additional heat treatments during the wafer manufacturing," said Manfred Bucher, senior manager of international marketing. "The second step was to refine the surface of the cost-effective substrate by depositing a very thin layer of defect-free silicon."
The target application for fLASH is advanced CMOS processes (at or below 0.18µm linewidths) with low well and junction depths. "The fLASH can also be used in DRAMs due to the high bit densities and it's particularly attractive to stacked trench technology," Bucher said.
At this time, the fLASH process is targeted at 200mm wafers, but Bucher expects that the technology will also be applicable for 300mm. M.V.
SC300 achieves functional 256Mb on 300mm wafers
After taking two weeks to convert its 300mm-wafer line from a capacity of 13 wafers/FOUP (front opening unified pod) to 25 wafers/FOUP, Semiconductor300 (SC300) announced that it has produced a functional 256Mb chip on a 300mm wafer.
"The 25 wafer FOUP was larger and required over 100 tools to be modified. The line conversion was completed in less than two weeks, which clearly demonstrates the capability and maturity of load port and FOUP supplies. This is a major step forward for productivity," said Horia Grecu, deputy general manager of SC300.
|
PHOTO. 13 wafers/FOUP and 25 wafers/FOUP at SC300 plant. (Photo courtesy of SC300)
The concept of the FOUP allows users to obtain maximum cleanliness inside the carrier at all times while the wafers are moved between tools, Grecu noted. This reduces the investment into additional air filtration systems when moving to smaller geometries. Utilizing >0.2µm technology, this milestone demonstrates the shrink potential of the trench-based process technology as well as the equipment capability for 300mm wafers. This news follows the recent sales of the world's first chips made on 300mm wafers.
Converting its tools was just the first phase, Grecu said; the second phase allows the plant to become fully automated. The plant still uses handlers, but expects to be fully automated in December with the equipment currently being installed.
SC300 will be using an automated system supplied by PRI Automation, Billerica, MA, said Bob Postle, VP and general manager of factory systems at PRI. The overhead monorail transportation system will use stockers, aerotracks, and hoists.
"The steps for full automation go as follows: wafers go from the equipment to the FOUP, then to the transport system, which takes the FOUP to the next process, then it goes to the load port," Grecu said.
SC300 is the joint venture between Infineon Technologies and Motorola. The plant produces 350, 300mm wafers/week. Grecu declined to discuss per wafer yields citing, "the products are sold by Infineon and all yield data is confidential."
However, earlier this year SC300 said it produced 64Mb DRAMs with pilot line yields of 60%. M.V.
X-ray stepper will have production role at Sanders GaAs facility
In what is likely the world's first application of "next-generation" lithography for revenue semiconductor production, Lockheed-Sanders has installed a Semiconductor Advanced Lithography (SAL) x-ray wafer stepper on its new 6-in. GaAs fab line in Nashua, NH, and will begin volume production of advanced microwave and mm-wave ICs (MMICs) with the tool in 2001.
The XRS 2000 stepper, which began making trial exposures Oct. 1, is easily capable of handling Sanders' current 0.15µm needs for gate levels, and should be able to migrate to geometries close to the 0.1µm regime that is in development at Sanders, said Bob Selzer, senior VP of technology at SAL, South Burlington, VT. A follow-on tool should be able to fabricate 70-80nm features within three to five years, he added. The system provides extremely wide depth of focus, in excess of 10µm. Less-demanding device layers will be fabbed on optical exposure tools.
Several important enablers for the program have emerged recently. One of the most important is the development of an x-ray point source by Science Research Laboratory, Somerville, MA. The NanoPulsar system provides enough x-ray radiation for calculated stepper throughput of 5-10 6-in wafers/ hour; this figure is expected to rise with the addition of a beam collimator and development of x-ray-specific resists (a deep UV resist is currently used).
Sanders' program also benefits from the recently announced collaboration between mask supplier Photronics Inc., the IBM mask development facility in Burlington, VT, and the Defense Department; the combined organization is providing masks for the Sanders effort on a commercial basis. The stepper also makes use of beamline components developed at the University of Wisconsin's CNTech synchrotron facility, and a Suss-developed spin-bake-coat module.
The NanoPulsar point source, in development for a number of years, uses a dense plasma focus pinch technique to generate 1nm x-ray radiation. An annular plasma is generated along a coaxial electrode structure in a neon environment using a set of eight tightly coordinated solid-state switching modules. The system is pulsed at 130Hz; as each pulse runs off the end of the electrode, it creates a pinch area in which neon atoms are stripped of all but one of their electrons, causing them to produce x-ray radiation with temperatures reaching the range of 6 x106 K in the pinch area. Published data and specifications indicate the system produces 2.6kW of spherical x-ray radiation at a repetition rate of 130Hz; a 50mj/sq.cm resist exposure takes 19 sec at rated power. The envisioned addition of a beam collimator will enable more radiation to be directed to the wafer being exposed, and reduce exposure time.
The source is compact enough to sit directly behind the stepper, and has been designed for rapid tool-free routine changeout of electrodes and anodes, said Dennis Reilly at SRL.
An important factor for the x-ray program is the relatively low feature density of the MMICs' gate-level masks. SAL president Dan Fleming explained that the reduced densities result in fewer defects, as well as faster mask repair times, thus enabling the mask house to achieve higher yields. P.D.
TECH BRIEFS
Ion Optics and the Jet Propulsion Laboratory have won an Advanced Technology Program (ATP) award from the National Institute of Standards and Technology (NIST). The two companies are joining forces to develop an integrated infrared gas "sensor-on-a-chip." This new chip integrates all the elements of a high-end industrial gas and chemical sensor onto a single integrated circuit that can be manufactured in high volume at a low cost. The project is expected to take three years.
Advanced Micro Devices (AMD) Fab 25, Austin, TX, said it has adopted Tokyo Electron Ltd's Unity DRM as its preferred oxide etcher for critical local interconnect applications involving cobalt silicide (CoSi). AMD migrated to CoSi from titanium silicide to achieve smaller gate designs.
Lucent Technologies and E Ink Corp. announced a joint development agreement that aims to move electronic books and newspapers resembling flexible plastic sheets one step closer to reality. The two companies plan to develop "electronic paper," which would be the first flexible, plastic electronic display entirely made with a process similar to ink-on-paper printing, rather than the more costly silicon-chip manufacturing process. Electronic ink and printed, plastic transistors are the key technologies.
null