Product category:
Electronics Manufacturing Services
News Release from: Applied Microengineering | Subject: MEMS devices
Edited by the Electronicstalk Editorial
Team on 25 February 2004
Time to rethink MEMS device manufacture
The design-house/foundry model for the fabrication of MEMS devices that his company has been pursuing for a number of years simply does not work, says AML's Rob Santilli.
Applied Microengineering (AML) has concluded that the design-house/foundry model for the fabrication of MEMS devices that it has been pursuing for a number of years does not work Basically the foundries' production processes are too inflexible for the fabrication of the wide variety of MEMS devices being proposed
This article was originally published on Electronicstalk on 27 Mar 2000 at 8.00am (UK)
Related stories
Anodic bonding joins silicon wafers to glass
AML launches its first, new, standalone, flexible, "off-the-shelf" Anodic Bonding machine the AML-400.
Wafer bonder improves MEMS gyro
The throughput and performance of wafer bonding tools can play a key role in determining the cost of assembly and packaging of MEMS and other microsystems devices.
In retrospect it's understandable - it is like asking the Mini car production line in Oxford to make coffee cups.
AML is reverting back its traditional in-house fabrication of MEMS chips with certain specialist processes being subcontracted out to AML-approved suppliers with which AML has built up relationships over the years.
The down side to this is that only small volumes can be satisfied and the "holy grail" of a flexible route to medium- and high-volume manufacture of MEMS devices without investing in your own facility remains elusive.
Further reading
Radical shift in wafer activation
AML's customers can finally activate, align and bond in-situ, in one chamber.
Nickel layer stops packages sprouting tin whiskers
Agere Systems reckons its engineers have found the right mix of packaging ingredients to enable the semiconductor industry to successfully implement lead-free packaging.
BGA inspection demands a closer look
Paul Cooper, Product Manager at Blundell Production Equipment, explains why visual inspection is preferable to X-ray analysis in evaluating BGA joints.
It seems the flawed concept is nobody's fault: it is simply just that established foundries are there to produce their own devices and have obviously developed and focused their manufacturing processes and process sequences to this aim.
The result is that they are necessarily rigid and are not flexible enough to satisfy the wide and varied MEMS device requirements.
AML has found that only 2% of MEMS requirements exposed to the route, might, be applicable.
For example, if you wanted to fabricate an accelerometer and essentially use the same process steps and sequences that your chosen foundry runs, you might have a viable option.
But this implies that you will try to make something very similar to what the foundry already produces.
Also there are probably enough pressure sensors and accelerometers on the market already.
A viable route to volume manufacture of MEMS devices without investing in your own facility remains elusive.
As does mimicking the IC industry with IC design houses able to work independently from IC foundries because all IC foundries run standard processes.
There are very few standard processes in MEMS, which makes this and second sourcing impossible.
This industry seems to be burying its head in the sand on this point, especially that of second sourcing.
For example, you could spend millions on the development of a device, get to production and the foundry with its unique processes burns down or goes bust, what then.
CMOS-compatible MEMS may have a brighter future because of standardisation, as ST Microelectronics and Analog Devices have shown, but CMOS MEMS will never be able to truly unleash the breadth of MEMS in terms of structures and materials, due to the restrictions CMOS processing applies.
And in any case, exploitation will be restricted to very high volume applications such as automotive, disposable medical and telecomms when it recovers.
AML can now only satisfy low volume MEMS device requirements by manufacturing them itself "in-house" and contracting out specialist processing such as glass micromachining.
Although this route should be aided and improved by the new DTI MNT network initiative which has just been announced by Lord Sainsbury.
Remember that MST, MEMS and MNT are manufacturing technologies.
There is no point in designing a device in isolation from where it will be made: the two are inextricably linked.
If someone is trying to entice you into a MEMS development without considering this, and identifying who and where production will take place are either trying just to get a contract out of you or they are incompetent.
Such mistakes can be very costly.
• Applied Microengineering: contact details and other news
• Email this article to a colleague
• Register for the free Electronicstalk email newsletter
• Electronicstalk Home Page
Search the Pro-Talk network of sites
