Silicon carbide and harsh environment MEMS applications

For high temperature or corrosive media environments, silicon carbide can be a material of choice and has the potential to open up new applications and markets.  We recently spoke with Professor Roya Maboudian of UC Berkeley about the history of silicon carbide, comparisons to silicon and other materials as well as current research and commercialization challenges and opportunities.

MEMS Investor Journal: What makes silicon carbide an advantageous material to be used in MEMS and sensors applications?

Professor Maboudian: In comparison to silicon, SiC has several key advantages. Its wide energy band gap, high thermal conductivity, large break down field, and high saturation velocity makes this material an ideal choice for high temperature, high power, and high voltage electronic devices. In addition, its physicochemical stability, high melting temperature, extreme hardness and wear resistance make SiC an attractive material for fabricating sensors and actuators that are capable of performing in harsh environments, such as high temperature, and corrosive and abrasive media.

MEMS Investor Journal: What are the disadvantages and limitations of SiC as compared to silicon?

Professor Maboudian: Although much progress has been made in making larger area SiC wafers, the quality is not as good as Si wafers and the cost is significantly higher.  As a thin film, the material, with tailored electrical and mechanical properties, is not widely available to the designers.   As far as MEMS is concerned, some of the on-going research opportunities include reducing the deposition temperature, tailoring the stress and strain gradient, developing selective etching schemes and tailoring the electrical properties. 

MEMS Investor Journal: How do development and volume production costs compare to silicon?

Professor Maboudian: Provided the unit process steps of thin film deposition, etching, and chemical-mechanical polishing of SiC are accessible, the die cost of a SiC MEMS device should be comparable to its silicon counterpart.  There will be up-front tooling and process qualification costs to establish these process capabilities in a commercial foundry environment.  If single crystal SiC electronic circuits are required (e.g., for certain high temperature applications), their cost will be significantly higher than silicon CMOS for the foreseeable future.  It is worth noting that the cost of packaging may be reduced, relative to silicon MEMS, if the SiC MEMS sensing element can be directly exposed to the harsh environment media.

MEMS Investor Journal: What are the competing technologies for harsh environment sensing applications?  How do they compare with SiC?

Professor Maboudian: From a material perspective, diamond has superior properties to SiC in many ways, with the exception that it oxidizes readily at high temperature.  However, diamond-based electronics lag significantly behind SiC electronics, as do deposition and related process technologies for cost-effective large-area thin films.  In other words, diamond MEMS is not mature enough for commercialization.  In many cases, SiC MEMS will compete with non-MEMS materials such as high performance metals, ceramics, and piezo-electric materials.  Each of these has its particular strengths and drawbacks, but in general macrosensors are costly and not easily miniaturized; common piezoelectric materials such as quartz are not well suited for high temperature operation.

MEMS Investor Journal: When did the first SiC MEMS application appear in research?  How did SiC R&D progress from there?

Professor Maboudian: For MEMS applications, the research dates back to early 1990’s, by Prof. Mehran Mehregany’s group at Case Western Reserve.  Since then, significant progress has been made in this field by his group and others including ours.  Initially, the Case effort was geared towards depositing SiC films on large area silicon substrates by hetero-epitaxy in an atmospheric pressure CVD reactor.  In the late 90’s, the LPCVD platform was established as a process more well-suited for commercialization and scalability.  Our work started in the late 90’s with the aim of reducing the deposition temperature and developing selective dry etching schemes.

MEMS Investor Journal: What are the main academic groups today which are doing research with SiC MEMS and sensor applications?

Professor Maboudian: In addition to CWRU and UC Berkeley, I’d like to mention work by Prof. Kornegay while he was at Cornell and by Prof. Steckl at University of Cincinnati.  There is also research activity in Europe, most notably by Prof. Cheung at University of Edinburgh and Prof. Sarro at Delft University of Technology.

MEMS Investor Journal: What are the current SiC related activities in your group?

Professor Maboudian: In partnership with CWRU, we are developing temperature, pressure, and acceleration sensors for harsh environments.  The aim is to make a complete SiC sensor including electronics that can survive high temperature, reactive environment and high g.

MEMS Investor Journal: Which startups are currently commercializing SiC MEMS technology?

Dr. Maboudian: FLXmicro is, to my knowledge, the first company to develop processes to cost-effectively microfabricate poly-SiC for high volume applications.

MEMS Investor Journal: Which large companies have SiC MEMS and sensors programs?  What kinds of products do they currently have on the market?

Professor Maboudian: Several companies have internal research in SiC.  What I have seen is that many companies in the past have looked at SiC for MEMS and sensors applications but due to the difficulties in processing this material, they have deemed it not cost-effective.  Now, with the advances made over the past few years, we are seeing that these companies are taking another look at SiC and seriously examining it for their applications.

MEMS Investor Journal: What are the main commercialization challenges with SiC applications today?

Professor Maboudian: As with any new technology, finding early adopters is important; there is often reluctance to change even if the new technology performs better and/or is less expensive than the current technology.  One key challenge has been the development of manufacturable process technologies for SiC MEMS that are cost-effective; this has, for the most part, now been demonstrated and is poised to be transferred to a commercial foundry environment.  Another challenge relates to the technical development required in the areas of signal conditioning electronics and packaging for harsh environments, i.e., the productization from a SiC MEMS chip to a sensor or actuator component that is qualified for a particular application. 

*******************
Roya Maboudian is a professor in the Department of Chemical Engineering at the University of California, Berkeley. She received her Ph.D. degree in Applied Physics from the California Institute of Technology in Pasadena, California. Prof. Maboudian's research interests are in the surface and materials engineering of semiconductor devices. Her most recent work has focused on tribological issues in micro- and nanoelectro-mechanical systems and development of novel processes for materials integration for high-performance MEMS/NEMS. She and her group have designed surface processes to reduce adhesion and friction in MEMS. More recently, they have developed new methods to integrate silicon carbide and diamond-like carbon films with MEMS technology. They are also developing schemes for selective deposition of high-performance metals in Si-based micromechanical devices. Prof. Maboudian is the recipient of several awards, including the Presidential Early Career Award for Scientists and Engineers (PECASE), NSF Young Investigator award, and the Beckman Young Investigator award.

Comments

[Neil Kane]

We appreciate Professor Maboudian's comments about the development of SiC MEMS applications. Prof. Maboudian correctly points out that SiC has exceptional properties which are exceeded only by diamond. But her characterization of the commercial state of diamond MEMS is not entirely accurate. Our company is shipping large area, foundry grade starter wafers made entirely of diamond for MEMS and sensor development. We agree with her characterization of the traditional impediments to developing MEMS devices with diamond, but those days are behind us. Diamond is now a viable MEMS material, and an affordable commercial supply of diamond substrates is available. More details can be found at our website: www.thindiamond.com

[art painting]

I took a quick first read through this and was quite impressed with what you have pulled together. I'm looking forward to going back and reading through it again with a highlighter and pen to make remarks and take notes.

Great stuff!