Standard Si processes drive economies of scale
05/01/2005
Solid State Technology invited executives from the compound semiconductor and MEMS industries to discuss how they are benefiting from economies of scale similar to those in traditional IC manufacturing.
IC industry: Showing opticom the way
 Ray Milano, Vitesse, Camarillo, California |
High-speed access and enterprise networks (whether fiber- or copper-based) have resulted in a new techno-economic environment for the development and delivery of optical communications products. The future trajectory for optical communications dictated by this new environment will be very similar to the revolution that occurred in the electronics industry where, arguably, the highest levels of innovation and systems complexity occur in the consumer electronics sector. Eblana Photonics has developed a new photon bandgap laser-technology platform that addresses the needs of this environment by enabling more useful optical components that can be manufactured more simply and cost-effectively using standard IC toolsets.
 James O’Gorman, Eblana Photonics, Dublin, Ireland |
In long-haul networks, the price elasticity of demand for optical components is relatively low, reflecting the relatively few optical components consumed by those systems and their low price when viewed in the context of the overall system. The reverse, however, is true in high-speed optical access networks that are now being deployed in volume around the world. In this new market, a high price elasticity of demand is the new norm and ultimately reflects the price that consumers are willing to pay for broadband access and their expectations for the services promised.
The underlying reasons for the similarity in evolution of the two industries are the same: It occurs when a technology’s further development and manufacture can only be profitable if there is a sufficiently large customer base over which the capital and development costs of a product can be amortized.
Fitness criteria for technology evolution can be summarized as follows: 1) the technology must deliver new and previously unobtainable and desirable experiences for large numbers of people; 2) it must be intrinsically manufacturable in high volumes; 3) it must be affordable; and, 4) it must be a platform technology with significant product extensions into the future with the characteristics of the first three. Finally, as performance price points are shattered, new uses and applications arise that reinforce this cycle by creating new high-value mass-market applications. (Examples of this are the way in which the disruption of the price points of visible HeNe and diode lasers enabled entire industries in barcode scanning and optical storage.)
Another key lesson from IC manufacturing is that the best way to reduce overall system cost is through higher-quality components with more uniform and repeatable characteristics. The result is a subsystem or system with high manufacturing yield and reliability; the converse is also painfully true.
A corollary to the previous statement is that IC integration proceeded gradually through a series of steps, the heights of which were purely determined by their ability to improve system performance while simultaneously reducing cost, improving manufacturability and yield, and either delivering or enabling a product or service that delights consumers by satisfying a higher unfulfilled need.
A key challenge for photonics and communications is to produce optical technologies that can be mass-manufactured with high electronics-like consistency and quality for mass-market subsystems or systems. These new approaches cannot represent a point solution in time to this problem, but must have an evolutionary path to more complex products. Older technologies developed for long-haul network applications, because of their price-inelastic evolutionary seed, are not capable of overcoming these challenges.
The photon bandgap laser technology solves a key problem in optoelectronics technology yet is manufactured using standard IC process tools. By using only standard electronics design rules and mature processes, it achieves performance and product consistency typical of IC products. To date, these have not been features of photonics products.
Electronic process line production of high-quality, higher-performance laser diodes has a significant impact on the communications industry. In the short term, it disrupts the production costs of photonic components by permitting low-cost manufacture of the highest-quality laser products on a mature electronics platform using only IC toolsets and processes and their associated process control regime. In the intermediate term, it will enable the required levels of photonics integration to drive the burgeoning rollout of fiber to the premises and gigabit Ethernet in access and enterprise communications applications. As in the microelectronics industry, integrating mature InP circuit technology with an InP laser technology will create new components and enable exciting new applications based on integrated photonics and electronics. The recent demonstration by Vitesse and Eblana Photonics is a key step in the journey.
Contact James O’Gorman, CEO, at Eblana Photonics, Dublin, Ireland; e-mail [email protected].
MEMS integration: A bridge to 3D IC manufacturing
 Henne van Heeren, Enabling M3, Dordrecht, The Netherlands |
The diverse nature of the MEMS industry makes it difficult to identify applications and technology trends that affect the industry as a whole. However, the merging of functions in products such as cameras, music players, computers, mobile phones, and portable consumer electronics in general, are specific applications where MEMS products must adhere to the consumer electronics industry’s quality and cost levels, and therefore provide an opportunity for change.
Microphones, lenses, and miniaturized data storage and energy supply units all benefit from MEMS manufacturing techniques and must adhere to the cost ethic used in the high-volume consumer electronics industry. Because MEMS products need to interact with external stimuli other than electronic signals, they are highly sensitive to the production environment. Technologies used in high-volume electronics with existing semiconductor supply-chain elements cannot always be adapted to the MEMS industry.
An example of this dilemma is found in the automotive sector. MEMS manufacturers need to use relatively flexible packaging technologies (e.g., ceramic and metal can packages). Companies delivering to the high-volume automotive market were forced to abandon these less cost-effective technologies in favor of adapted plastic packages (i.e., using an optical window or a product-specific cap). This step was seen as a major breakthrough in the cost/quality debate that manufacturers struggled with for many years.
Continuing to enter the consumer market requires additional steps. The MEMS industry is now embracing advanced technologies, such as wafer-scale packaging, to achieve even lower costs and less usage of real estate on the wafer and PCB. It also avoids device damage during processing. Wafer-scale packaging for MEMS will be introduced first for inertial sensors and fluidic devices. Fluidic devices tend to combine wafer-scale packaging with electronics and/or sensing functions onto the silicon wafer with fluidic functioning in the capping wafer. Until recently, the use of wafer-scale packaging in MEMS was limited by the need for high temperatures and/or high voltages to achieve secure bonding. This not only limited the implementation to a few (not very sensitive) devices, but it also had serious limitations in terms of cycle time and efficient equipment use.
Recent developments in low-temperature bonding, however, have led to more cost-effective bonding processes. A process for small wafer through-holes, currently under development by a leading MEMS dry reactive-ion etch (RIE) equipment supplier and one of the largest lithography suppliers, paves the way to simple through-wafer interconnection concepts. What is missing is a supply chain able to sustain these technologies, as well as a more standardized layout of capping wafers to enable the initial low volume that MEMS component suppliers will have while developing this technology. A standardized capping-wafer layout could enable production facilities in the backend.
Another factor to consider in using wafer-scale packaging for MEMS-based devices is the availability of packaged-product reliability test results to confirm the validity of these technologies.
Following are other examples in which the adoption of MEMS technologies by IC factories for the fabrication of monolithic MEMS devices and advanced semiconductor products has forced them to become more compatible to semiconductor processing.
- Deep RIE: The etch rate has increased to meet cost expectations in the IC industry.
- Backside alignment is becoming available on standard steppers.
- IC-compatible TMAH etch is replacing anisotropic silicon wet etch.
- LIGA (lithography-electroforming-replication) and HARMST (high aspect-ratio microsystem technology) are alternatives using UV-sensitive positive resist formulations.
- Sacrificial etching: Available dry etch technologies are omitting yield-sensitive wet processing.
- Release etch: CO2 drying.
MEMS EDA tool suppliers also have developed software to design and simulate what is simply a 3D process. Much work remains to optimize the processes to achieve low equipment cost-of-ownership and high yields. This will not only lead to more monolithic MEMS products, but it will also help move the semiconductor industry into the 3D world, where additional IC functions are integrated in the third package dimension.
Continuing to adapt MEMS manufacturing technologies for compatibility with standard IC manufacturing will enable MEMS to transform from a low-volume industry primarily serving market niches into an important part of the high-volume electronics industry. It will require an active, innovative community of service suppliers providing equipment, software, and services to optimize knowledge and capacity.
Contact Henne van Heeren, Enabling M3 business development manager; e-mail [email protected].