Processes for MEMS by Inkjet Enhanced Technologies
Printed electronics achieved recently considerable progress due to new printing technologies and to the introduction of nanoparticle inks, paving the way towards integrating its capabilities within the silicon-based nanoelectronics. The objective of the ENIAC JU project PROMINENT was to demonstrate significant cost reduction in MEMS manufacturing by using printing technologies to reduce materials, chemicals and energy consumption, waste water production, processing cycle time and capital investments.
MEMS technologies experienced a fast development during the recent years. Europe has globally a strong position in MEMS manufacturing, but in order to maintain its leadership it must significantly reduce costs. The ENIAC JU project Prominent was planned to develop novel low-cost, digitally controlled additive manufacturing methods that could radically change the manufacturing methods for MEMS and bring a substantial competitive edge to the European MEMS industry. The objective was not to replace the whole MEMS manufacturing process, but rather to introduce a novel flow – more flexible and cost-efficient – by using methods developed in printed electronics.
Work and consortium
Integrating printed and silicon-based electronics has only become possible recently, due to the new developments in printing technologies and the introduction of nanoparticle inks. An integrated approach would draw benefits from both technologies.
Additive deposition has a large cost saving potential and straightforward applicability in vertical interconnections and through-silicon vias where it could clearly simplify the process. As an alternative, low cost laser machined vias for vertical interconnections and inkjet-assisted methods were planned to be used for via structure creation. In inkjet filling, a critical challenge is to find a suitable combination of materials and processes to solve the wetting problem perceived with nanoparticle fluids.
Sometimes functional materials are needed within the hermetic cavity of the MEMS device, e.g. for controlling the gas pressure inside the device. The materials used for this purpose are conventionally deposited on the capping wafer by using subtractive methods, but the cap wafer topography may make that difficult. Cost savings can be obtained by using additive methods. As a demonstrator, the current getter deposition process was planned to be replaced by the new one in modified MEMS inertial sensors and / or timing devices. Wafer bonding with inkjet printing were also to be studied. Surface treatment methods such as local plasma processing – as well as conductive and convective heat sintering – were predicted to ensure the printed bonding layer seals the MEMS cavity hermetically.
The multi-disciplinary Prominent required expertise in different technology areas; ranging from MEMS research and wafer manufacturing to inkjet printing machinery. Thus, the consortium was built of a number of leading experts from universities, small and medium size companies, and large MEMS manufacturers, representing different value chain levels of the MEMS industry. The parties had complementary expertise in conception, design, evaluation and manufacturing of MEMS components, as well as printing and surface treatment equipment development and manufacturing
It is essential that new alternative methods are integrated within existing manufacturing processes without compromising device quality, reliability, yield or throughput. Quantitative methods for detailed assessments of cost reduction, environmental and market impact were to be developed.
In particular, executing selected steps in the MEMS manufacturing using maskless, digitally controlled, localized additive processes instead of the incumbent subtractive processes were predicted to result in a greatly simplified process sequence. This was predicted to result in
- lower initial investment costs for a MEMS line, making it easier for manufacturers to introduce new products
- new features in the MEMS devices, new application areas
- increased flexibility in production, allowing for smaller batches, mass customization and fast changes in the production process
- easier prototyping and shorter time-to-market period in new MEMS devices
- greatly reduced production costs and environmental impact
Results achieved during the project periods
During the first 12M period initial specifications were defined, different inks are tested for both TSV, RDL and bonding structures, and selective surface activation is used in development. One of the partners has acquired a novel inkjet system enabling superfine structures earlier unachievable with conventional technology and the system in now implemented in work. Also first 3D IC-MEMS assembles have been realized.
During the second reporting period, ending at M24, all potential ink-jet material pre-evaluations were done and most of the testing completed and based on these results final demonstrators were selected. Surface activation for inkjet process was successfully demonstrated and Innophysics introduced a new configuration of the µPlasmaPrint system to commercial markets. Partner poLight announced during the period collaboration with STMicroelectronics to ”use STMicroelectronics TFP MEMS technology to manufacture its TLens autofocus lens, a product that brings new camera autofocus performance to smartphones”.
During the third period the focus was laid on to complete the demonstrators. Notable achievements were attained in ink-jet deposition of protective layers in MEMS fabrication, conductive interconnects between layers, RDLs, and self-assembling TSVs as well as techniques of gettering and realization of controlled pressures in structures. A detailed cost analysis and a tailorable software tool were finished to compare different process flows and cost differences of the ink-jet processes compared to more conventional techniques.
FO-WLP packaging of two different kinds of functional MEMS devices, one having an opening to ambient, was demonstrated. With the method low-cost very-thin MEMS devices can be made in volumes without being limited to completely sealed packages. Furthermore, MEMS autofocus lens development was successfully completed and demonstrated. This device has faster focusing speed and lower energy consumption compared to existing devices in the market.
Ink-jet technology offers attractive possibilities on MEMS manufacturing being additive, maskless process with potential cost savings. However, ink formulations play a crucial role, in each application there are some unique requirements, like purity of the deposited material after curing or stability of the material.
Overall results achieved during the whole course of the project
After the three years and four months’ project, as a general observation and as one of the results of the project it can be summarised that ink formulations influence strongly the ink-jet printing results, and in each application there are several unique requirements, possibly making standard off-the-shelf inks unsuitable for the desired purposes or their performances are not necessarily at the optimum. In other words, inks (and stability of nanoparticles), especially with reactive materials, like silicon, titanium or copper, need further development. Despite of this, (i) RDL (redistribution lines) applicable to MEMS devices with ink-jet printing, (ii) ink-based interconnects in SOI wafers (related patent application pending), (iii) wafer-to-wafer bond frames deposited with metal inks, and (iv) additive etch mask manufacturing avoiding normal lithography steps were all demonstrated.
Some of these results can after minor development efforts be transferred into production especially now when – unlike during the planning phase of the project – ink-jet tools capable of handling silicon wafers cassette-to-cassette, are available. The tools also accept the plasma activation head introduced and developed by partner Innophysics in the project. Innophysics’ device can modify wafer surfaces locally in order to influence the wettability of the ink at the surface. This way it is possible to draw very fine ink lines on silicon, and it is also possible to fill cavities or TSV´s without over-spilling the ink.
Partner Silex demonstrated together with partner KTH novel TSV and TGV concepts. In addition, Silex demonstrated a gettering process suitable for MEMS devices, and, invented and developed a novel method for multiple pressures suitable for WLP-based MEMS packaging for example for combo sensors. This beyond the state-of-the-art technology was validated by Silex with realistic test structures giving a pressure difference up to 4 orders of magnitude in a same MEMS die (1 Pa vs 104 Pa). A patent application was filed by Silex in 2014, and was granted just before the end of the project. Silex developed a new wafer level capping technology for RF MEMS using innovative novel TGV in glass showing very high linearity and low RF losses, and during 2016 Silex started to commercialize it.
To our best understanding, in this project partners Murata and Nanium demonstrated the very first FO-WLP packaged inertial and pressure sensors. FO-WLP offers ways to make very thin, < 0.5...0.6 mm of thickness, sensor packages. Furthermore, in the project, Murata introduced a piezo-actuated inertial sensor platform successfully.
Partner poLight developed and demonstrated TLens, a novel piezo-actuated autofocus lens for mobile communication. The device is faster than conventional designs, and also the current consumption is greatly reduced compared to existing solutions. The implementation of the lens, for commercial production at a foundry, was done in other Eniac JU project (Lab4MEMS) successfully. In this project NANIUM designed a FO-WLP Process Flow for hollow dies integration.
As a summary, most far-reaching targets were achieved partially, and need further development and close collaboration with nanoparticle and ink manufacturers, but a considerable part of the results is applicable in commercial production while some results are even ground-breaking achievements.
The net cost effect will include reduction of materials, chemicals and waste water, used energy and processing time, and the capital investments of setting up a new process line.
Key project dates
- Start: 1.3.2013
- Finish: 30.6.2016
- The Netherlands
- € 9.4 million