Semiconductor device manufacturing: process – Chemical etching – Vapor phase etching
Reexamination Certificate
2000-10-27
2003-03-18
Utech, Benjamin L. (Department: 1765)
Semiconductor device manufacturing: process
Chemical etching
Vapor phase etching
C438S711000, C438S720000, C438S722000
Reexamination Certificate
active
06534413
ABSTRACT:
CROSS-REFERENCE TO RELATED APPLICATIONS
Not applicable.
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
Not applicable.
BACKGROUND OF THE INVENTION
The present invention is directed to fabrication of micro-electromechanical systems (MEMS). In particular, the present invention is directed to removal of a silicon dioxide sacrificial layer during the fabrication of the MEMS.
MEMS utilize movable structures to perform their function. Examples of movable structures include cantilever beams (accelerometers), diaphragms (pressure sensors and microvalves), freely rotating structures (micromotors and microgears), or hollow tubes suspended over a cavity (micro-densitometer). The fabrication of these structures often incorporates a layer of silicon dioxide upon which the structure is fabricated. At some later step, the silicon dioxide layer is removed, thus freeing or releasing the structure. The silicon dioxide in this instance is an example of what is commonly referred to as a sacrificial layer. Silicon dioxide is the most common material used as a sacrificial layer. These layers are commonly removed with a liquid-based release etch.
There are two problems associated with the release etch. First, after the liquid-based release etch, it is not uncommon for the freed structure to be permanently adhered to the wafer, rendering the device useless. One reason this happens is that the liquid etchant can hold the freed structure in close contact to the substrate through capillary forces. As the liquid evaporates, the structure can become permanently affixed to the substrate either by bridging the gap with residue from the etchant or by “stiction”. Stiction is a term describing the force of adhesion that exists between two surfaces. Even clean, smooth surfaces can have strong adhesive forces. Thus, stiction is not simply a matter of surface cleanliness.
The second problem is that the release etch tends to be incompatible with the steps required for fabricating microelectronic circuits, especially CMOS devices, unless special precautions are taken. This situation has contributed to the slow progress in merging MEMS devices with on-chip electronics. At present, such integration is achieved only if the circuitry can be completed before the release etch is performed, or if the MEMS device can be completed and encapsulated prior to the fabrication of the circuit. However, it is very difficult in practice to achieve such a design. Microelectronic processing steps require very clean surfaces, especially free of metallic contamination. MEMS processing steps, on the other hand, tend to be relatively dirty, with high levels of metal and organic contamination. Normally, in microelectronic processing steps, the wafer would be cleaned using a liquid phase clean. However, for MEMS devices, a liquid phase clean can lead to stiction problems for the freed microstructure.
There are also other instances where one might wish to keep a surface free of metal contamination while also removing sacrificial silicon dioxide. For example, on-chip cell testing devices in the area of BioMEMS (MEMS devices with a biological function) or devices intended to be used as chronic implants into biological systems (e.g. humans).
Integrating the micromechanical and the microelectronic devices onto a single chip has been an objective of the industry since the first MEMS devices were made. The initial challenge was to make micromechanical devices using the same or similar fabrication steps as microelectronic processing, thereby capitalizing on the economy of scale associated with mass production. MEMS devices are now being developed under guidelines that are considered CMOS compatible, which means that the devices are fabricated using only those processing steps and materials that are available for fabricating CMOS devices (i.e. silicon; silicon oxide; metal, typically aluminum and tungsten; and photoresist). Clearly, this imposes a stringent limit on the techniques and materials available to the design of the integrated systems. The circuits would be responsible, for example, for controlling the MEMS device, interpreting a signal from the device, and transmitting the needed information between the MEMS device and the macroscopic world.
One approach to making the integration of the MEMS and CMOS technologies possible is to perform the release etch at a stage after the electronic circuitry has been completed and, if required, a passivation layer deposited over the sensitive area to protect it from the release etch (see, for example, Dai, Ching-Liang and Pei-Zen Chang, “A CMOS surface micromachined pressure sensor,”
Journal of the Chinese Institute of Engineers
, 22(3), pgs. 375-80 (1999); Waelti, M., et al., “Package quality testing using integrated pressure sensor,”
Proceedings of the SPIE—The International Society for Optical Engineering
3582, pgs. 981-86 (1999); Scheiter, T., et al., “Full integration of a pressure-sensor system into a standard BiCMOS process,”
Sensors and Actuators A
67, pgs. 211-214 (1998); Buhler, Johannes, Steiner, Franz-Peter, and Baltes, Henry, “Linear array of CMOS double pass metal micromirors,”
Proceedings of the SPIE—The International Society for Optical Engineering
2881, pgs.75-82 (1996); Stadler, S. and Ajmera, P. K., “Integrated acceleration sensors compatible with the standard CMOS fabrication process,”
Proceedings of the SPIE—The International Society for Optical Engineering
, 2649, p.95-100 (1995). Though there are a wide array of simple systems that can be constructed with this approach, it is limited in several ways. First, it requires that the electronic device and the micromechanical device be mostly independent of each other in a physical sense, interacting with one another only by way of an exchange of electronic or optical signals. Second, the processing of the integrated devices must be such that the electronic device is completed before the release etch is performed for the micromechanical device.
An alternative approach is to complete the MEMS device first and then build the microelectronic circuit. See, e.g., U.S. Pat. No. 5,798,283 (Montague et al.) and U.S. Pat. No. 5,963,788 (Barron et al.). In this case, the MEMS device must be encapsulated to prevent downstream contamination of the process tools needed for the microelectronic processing steps. This approach has drawbacks similar to those mentioned previously. In either case, the level of integration is relatively low in that the two devices are simply on the same chip and can exchange some type of signal. A higher level of integration would be to implement the electronic circuit onto the MEMS device.
At this level of integration, it may be necessary to perform the sacrificial etch prior to completing the electronic device. The sacrificial etch would release a structure or create some type of cavity to provide, for example, thermal isolation of a device. Such an ability would be imperative if it was desired to place the electronic device in the cavity or between two moving structures. However, the wafer would require cleaning after etching in order to begin or continue the microelectronic processing steps. The industry standard clean in this case would be a derivative of the RCA clean, which is a liquid phase process.
Liquid release etchants have the problem of leaving behind residue and, more significantly, leading to the permanent adherence of the movable structure to the fixed surface. When the sacrificial layer is silicon oxide, or more frequently phospho-silicate glass, the preferred etchant is generally either aqueous HF or buffered HF. See, e.g., Buhler, Johannes., Steiner, Franz-Peter., and Baltes, Henry, “Silicon dioxide sacrificial layer etching in surface micromachining,”
J. Micromech. Microeng
, 7, p. R1-R13 (1997). Clearly, other materials may be suitable for the sacrificial layer and other options are available for the release etchant, but whenever there is a liquid phase present, there is a significant potential for stiction to become a problem.
The semiconductor industry has been experimentin
Beck Scott Edward
Robertson, III Eric Anthony
Air Products and Chemicals Inc.
Brown Charlotte A.
Chase Geoffrey L.
Utech Benjamin L.
LandOfFree
Method to remove metal and silicon oxide during gas-phase... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method to remove metal and silicon oxide during gas-phase..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method to remove metal and silicon oxide during gas-phase... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3042850