With the recent introduction of MEMS in consumer applications (mobile phone, gaming…) the market for low cost and low power consumption multi-axis inertial sensors has boomed over the last five years. Moreover, complex motion features are now required thus leading to use not only one modality of sensor (like accelerometer only) but also combos (6-axis or 9-axis sensors, combining accelerometers and gyroscopes, accelerometers and magnetometers, or all sensors simultaneously). These sensors are used for motion sensing in various applications such as handheld device (remote control, mice, mobile phone, ...) for user interface or man-machine interactions, physical activity monitoring, situational awareness (power management, feature activation, hard disk drive protection), image stabilization for camera phones (pitch and roll motion detection), localization (to provide dead reckoning when GPS signals are lost), rehabilitation, …
To correctly accompany the huge and fast progression of the MEMS market in consumer applications, three major trends have to be addressed:
· The size reduction directly linked to the cost of the device
· A low power operation, especially for the gyroscope
· The integration of combos (several sensors, same die) to accurately solve the motion estimation problem.
Today, the European industry has the world leadership on the MEMS consumer market with a high expertise in design of both sensor and associated electronics. But, to keep this leadership, it has to already think to the next move. To fill the very aggressive requirements of the next generation, a simple optimization of design and process of existing concept will certainly not be sufficient. A new technological breakthrough is then required to reach a higher level of integration and reduce the manufacturing cost. This is clearly the first main objective of NIRVANA project: to propose an innovative sensor concept and technology based on the use of nano-scale detection means, which allows both, a high degree of miniaturization and a full integration of a low cost 9-axis inertial sensor for accurate motion sensing.
The need for very small and very low power consumption inertial sensors is also emerging in high-end markets and especially the medical one. The most known of MEMS medical application is the cardiac pacemaker but others come out such as the rehabilitation of severe balance dysfunction using functional vestibular prosthesis or hearing impairment using cochlea and middle ear implants. For implantable prostheses, the low power consumption is a key parameter as it directly affects the battery lifetime and the battery replacement frequency which often requires invasive surgery. This leads to the second objective of the NIRVANA project: to operate optimally, in terms of mechanical and electronic design, all the properties of nano-silicon gauges to decrease very significantly the overall consumption of these inertial sensors.
To answer these objectives, a new design and a new detection mode have been investigated, thus giving birth to the “M&NEMS” concept. The basic idea is to combine on a same device a thick MEMS layer for the inertial mass with a thin and narrow NEMS part as suspended strain piezoresistive gauge. This concept involves two amplification effects:
· The inertial force on the seismic mass due to acceleration is concentrated on the nanometric section of the gauge resulting in a high stress applied to the gauge. Thus, the seismic mass size required for delivering the nominal stress in the gauge can be very small.
· The design of the structure includes a lever arm effect : the force applied to the gauge can be one or two order of magnitude higher than the inertial force on the seismic mass
Another advantage of the concept is the very high efficiency of the transduction: the only stressed part of the structure is the detection element, in this case the nanogauge and the whole produced stress participates to the piezoresistive effect and consequently to the signal.
This concept can be applied to in-plane and out-of-plane accelerometers and gyrometers through the same manufacturing process. The magnetometer requires additional steps for the integration of a permanent magnet on top of the structure. Compared to other magnetometer concepts, (Hall effect, AMR, GMR, Laplace force), the use of permanent magnet for the magneto mechanical transduction leads to a lower power consumption. With the choice of exchange-bias coupled antiferromagnetic and ferromagnetic multilayers patterned as perpendicular oriented stripes, the two magnetization directions required for a 3D magnetometer can be achieved. The development of this material is included in the project.
Thus, with the M&NEMS concept, the entire 9 axis can be integrated on a same chip in a compact and low power consumption package, theoretically filling the two main objectives of the project. This capability will be assessed on two demonstrators:
· The first one is focused on a miniaturized 3-axis gyrometer, integrated on a single die, with very low power consumption, for a medical application: a functional vestibular implant.
· The second one is a highly miniaturized 9-axis inertial sensor for consumer application. The goal is to integrate the 9-axis on a single chip, putting emphasis on the low power consumption.
The achievement of these demonstrators is divided in two steps:
· The first one is dedicated to the validation of each function independently (3A, 3G and 3M on separate chips associated to the corresponding ASIC)
· The second step consists in an optimization of the designs through the experimental results got on the first step demonstrators and in the realization of the two final demonstrators.
To reach the objectives, the main technical issues to address are:
· The 9-axis sensor electromechanical and ASIC design to fit the end-user specifications
· The technological developments including the integration of the silicon nanowires gauges, the integration of the magnetic material, the development of through silicon via (TSV) and copper pillar, of a low temperature packaging
· The 3D integration, from wire bonding interconnection for the first step demonstrators to copper pillar interconnection for the final demonstrator
· The characterization and reliability of inertial sensor
To ensure the success of this ambitious project, a motivated and excellence consortium has been established, with leading research groups (LETI, POLIMI, FhG-IIS) and key industrials partners: two end-users (MEDEL and MOVEA), and the inertial MEMS manufacturer world leader in the consumer segment (STM), will guarantee the real and prompt exploitation of the developments and results of the NIRVANA project.
During the first year of the project, the specifications of the final demonstrators were defined by the end users, STM, MOVEA and MEDEL. The consumer specifications given by MOVEA target the 3D orientation (when computed with a nine-axis IMU) and the pedestrian navigation. This latter feature is a challenge and requires very accurate heading estimates. The medical specifications for vestibular implants, given by MEDEL, differs from the consumer application in such a way that they are more drastic regarding the low power consumption while size requirements are relaxed.
In strong link with ASIC designers, electromechanical modelling and design of the MEMS have been performed to meet the specifications, on one hand by POLIMI for the gyros and on the other hand by LETI for the accelerometer and magnetometer. In particular, design optimization versus the nanogauge geometry and MEMS thickness were carried out. Original gyro designs have been proposed by POLIMI.
A draft Design Rule Manual and Design Rule Checking were first elaborated by LETI for POLIMI to be able to start the layout design and evolved to finalized versions through a lot of exchanges between POLIMI and LETI teams.
A first ASIC version was designed by the IIS institute with continuous CAD and design kit support of STM. The goal is to drive and provide a digitalized readout for the nano-gauge based sensors with very low power consumption. Separate conceptions for the gyrometer and the accelerometer/magnetometer using the STM 0.13 µm technology node have been implemented. Due to the high voltage requirement of the excitation circuit of the gyrometer it was necessary to select the high voltage process HCMOS9A from STM.
In parallel, the architecture of the first demonstrators was defined, including the footprint of the encapsulated MEMS, the number, location and definition of the MEMS and ASIC pads, the respective arrangement of the MEMS and the ASIC in the ceramic package.
The second year of the project has been mainly dedicated to the demonstrator fabrication: the MEMS on the 200mm CEA facilities and the ASIC at STM. Concerning the MEMS fabrication, difficulties have been met for the implementation of the wafer level packaging based on an eutectic reaction between gold and silicon to hermetically bond the cap wafer and the device wafer. The delivery of packaged MEMS has been postponed. Nevertheless the first characterizations of non packaged MEMS have been performed. Fatigue tests have showed that nanostructures can withstand high level stress during several years. Preliminary tests on sensitive structures have shown a good agreement with theoretical predictions. In parallel, the first tests of both 3A3M and 3G ASIC have been done allowing the validation of the functionality of the test board and the availability of all the functions and sub-modules. The architecture of the demonstrator 2 was also defined: the estimation of the size of the sensitive device combined with design rules constraints led to a size of the LGA plastic packaged 9-axis sensor of 4x4mm˛
Regarding the technological developments, the main difficulties of the process have been addressed (nanogauge realization, increase of the epitaxy thickness, mechanical structure etching, …). The development of the magnetic material (NiMn/CoFe) made of an alternation of coupled CoFe ferromagnetic layers and NiMn antiferromagnetic layers is in progress. An experimental design including various parameter screening have highlighted the major trends and shown the way towards the elaboration of magnetic layers with optimized properties. The integration of this material in the M&NEMS process flow is on-going and good results were obtained.
The hermeticity of the wafer level packaging is an issue. The retained technology was a gold-silicon eutectic bonding. Experiments on dummy wafers showed a good mechanical strength. Hermeticity will be tested measuring the quality factor of accelerometers encapsulated under vacuum.
The signal processing of the final demonstrators is anticipated through the development of a software platform to demonstrate Motion Features and prepare future inputs for Nirvana sensors. Calibration procedure for multi-axis sensors is also in progress
The expected final results and their potential impact and use (including the socio-economic impact and the wider societal implications of the project so far),
The expected final results of the project are the realisation of the 2 demonstrators described above with characteristics that meet the specifications defined at the beginning of the project. It includes the MEMS and ASIC design and manufacturing, their characterization, the test of their reliability, their assembly and final test in the environment of the targeted application. The success of the project is determined by a strong involvement of industry participants interacting closely with R&D organisations and users.
It is expected that this new concept will have short term applications in consumer markets and will contribute to the competitiveness of the European industry. The technological breakthrough of this new concept could contribute to keep inside Europe the leadership in motion MEMS business and maintain STM leadership versus US companies like Invensense or Analog Device, or versus Asian emerging new player like TSMC.
With the M&NEMS concept, there is potential for the medical device industry to gain low power miniaturized gyroscopes, accelerometers, magnetometers and combinations thereof enabling new medical applications and progress and gain considerable competitive advantage and economic growth for European medical device manufacturers. Since yearly production quantities are magnitudes lower than e.g. the consumer market the medical device industry typically cannot afford the multimillion technology investments needed for MEMS. With NIRVANA project funded at European level there is potential that special customized versions for the medical device industry can be spin off at a cost level affordable for the targeted application and products.
NIRVANA project applications concerns two fields of strategic importance for the European society. One is the performance improvement and functionality increase of consumer electronic devices used in everyday life which integrate human-machine interfaces such as gaming console, smart phones, … Developments can also find applications in image stabilization of digital cameras, hard disk drive safety, .... The other, and not the least, is related to health. One of the main objective of the Nirvana project is the reduction of power consumption which allows the emergence of new implantable devices that can be used, for example, to rehabilitate severe bilateral vestibular sensory loss or hearing impairment. These diseases affect several million people in Europe especially elderly or fragile people and has a considerable detrimental effect on the quality of life. Then, the expected results of NIRVANA will not only be of great scientific and economic interest but can have a huge impact on the society needs.
Address of the NIRVANA website: http://www.nirvana-fp7.eu
To address all the needs described above, the NIRVANA project aims at develop ing a very low cost and very low power consumption 9-axis inertial sensor based on a new concept and technology using nano-scale detection means. The idea is to take advantage of the very high sensitivity and low impedance of silicon nano-wire gauge to optimize the dimensions and the integration of the sensor and to reduce the overall power consumption.
At this end, innovative sensor designs, disruptive technology (mixing MEMS and NEMS technologies, and including packaging with TSV), and new electronics architectures will be developed in this project. The realistic goal we target is to fully validate the integration of 9-axis sensor based on nano-wire gauge detection, having a surface and cost 2 to 4 times less than the current commercial components and with power consumption 3 to 5 times lower. Special emphasis will also be put on the characterization and reliability of these sensors.
To ensure the success of this ambitious project, a motivated and excellence consortium has been established, with leading research groups (LETI, POLIMI, FhG-IIS) and key industrials partners: Two end-users (MEDEL and MOVEA), and the inertial MEMS manufacturer world leader in the consumer segment (STM), will guarantee the real and prompt exploitation of the developments and results of the NIRVANA project.
|March - 2019|