Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1.1 Motivation and Context for Research

Microelectromechanical systems (MEMS) originally referred to microscale systems made of both electrical and mechanical moving parts, being able to perform something interesting or useful. They are μ-sized 3D structures, realized with lithographic processes, which can interact with external world and they are typically integrated with electronic circuits. The presence of mechanical moving parts let them be sensors and actuators of our environment; they are reliable and cost effective.

In 1959 the Nobel Laureate Richard Feynman gave his inspiring talk “There’s Plenty of Room at the Bottom,” inviting the scientific community to exploit the huge amount of opportunities offered by the unexplored field of micro- and nano-fabrication and emphasizing the possibility of manipulating matter on atomic scale. This talk may be considered to some extent the birth or, at least, the precursor of MEMS world and evolution [1]. Around the end of the 1970s first micromachined devices were developed by different companies, like Hewlett-Packard (HP), Analog Devices (ADI), and Texas Instruments (TI). HP was the first to develop a proprietary technology to manufacture inkjet cartridge, TI developed digital micromirrors for light manipulation, and ADI is commonly recognized as the pioneer for MEMS accelerometers. European companies like Robert Bosch GmbH and STMicroelectronics played an important role starting at the beginning of the 1990s—Bosch process for Deep Reactive Ion Etching (DRIE) was patented in 1994—and they are still important players on the global market for different types of sensors.

Nowadays, MEMS are an expanding technology involving many different markets such as automotive (e.g., tires pressure monitoring systems, airbag systems, ESP, self-driving capability, etc.), consumer (e.g., gaming, navigation, image stabilization, augmented reality, etc.), home automation (e.g., distance speech interaction, powering wireless sensors where solar energy is not available, etc.), healthcare (drug delivery, cardiac monitoring, implantable neurostimulation, air quality monitoring, etc.), and industrial (building stability, structural monitoring, etc.).

While, up to few years ago, MEMS were thought as single sensor or actuator, the present trend sees MEMS to be a combination of functions. Inertial Measurement Units (IMUs) are multi-parameter systems in one single package based on multi-axis MEMS and they play a fundamental role in several fields of applications like navigation units in smart mobile phones and other consumer electronic goods [2]. The realization of an inertial measurement unit is of great interest for civil and military aviation, space satellites, trains, ships, unmanned or remote operated vehicles, stabilization systems, consumer electronics, and several other applications [3, 4]. In particular, as far as consumer applications are concerned, an increasing number of smartphones and mobile devices provides users with location- and orientation-based services in addition to “more traditional” functionalities based on navigation. Navigation is becoming a must-have feature in portable devices and the presence of a compass also makes location-based augmented reality emerge: a street map or a camera image could be overlaid with highly detailed information about what is in front of the user. To make these features possible both industries and scientific research are focusing on three-axes magnetic field sensors (also called magnetometers). And, like the rest of electronics industry, compasses are required to scale down, get cheaper, and more energy efficient [5, 6].

In this scenario, the integration of a three-axis accelerometer, a three-axis gyroscope, a three-axis magnetometer, and a pressure sensor, all based on the same MEMS process, can result in a “10-degrees of freedom (DOF)” high resolution, low-cost, and low-power miniaturized system for position, motion, heading, and altitude monitoring, representing a challenge for research. According to recent market reports, fusion of sensors is starting to be widely used in consumer fields and inertial MEMS market is booming and expected to register a CAGR growth of about 7. 8% by 2020 [7].

The state of the art for IMUs integrated using MEMS processes only is represented by a 6-axis MEMS unit for acceleration and angular rate sensing [8, 9]. Dominant technologies used to integrate magnetic sensing elements in IMUs which are currently on the marketFootnote 1 are not based on micromachining processes yet. The most common implementations of electronic compasses are based on Hall effect and on anisotropic magneto-resistance (AMR) technologies. For instance, the multi-axes systems for motion and magnetic field sensing described in [10, 11] include accelerometers or gyroscopes based on MEMS technology but magnetometers based on AMR [12].

Thus, AMR and Hall effect technologies still represent the most common implementations of electronic compasses. Devices based on AMR technology require the deposition of specialized magnetic materials into standard industrial processes, they suffer from magnetic hysteresis, require to be continually degaussed and, finally, they are sensitive only to in-plane fields. The magnetometer for the Z-axis is the same as for X and Y axis but in order to integrate a three-axis AMR-based sensing system into an IMU, a vertical assembling for the Z-axis device is required, which has a strong impact on the packaging cost and it represents an obstacle for ultra-thin packages. Large volume is then wasted in height due to the required out-of-plane assembling, reducing the integrability of the IMU in ultra-thin plastic packages. Even though Silicon Hall effect sensors avoid introduced problems arising from the integration of magnetic materials, their power consumption and resolution performance are relatively limited with respect to AMR devices.

The research on an alternative approach, Lorentz force-based MEMS magnetometers (in particular for the Z-axis element), is encouraged by the possibility to integrate magnetic sensors in the same MEMS technology already used for gyroscopes and accelerometers, so to design a complete 9-axis IMU in a single standard process. Moreover, among advantages of the integration of compasses in MEMS technology are sufficiently high resolution, sufficiently low power consumption, and no need to use ferromagnetic materials. In a very interesting intermediate step, a 9-DOF IMU can be realized with a 2-axis AMR magnetometer and a 7-DOF MEMS die, with a consequent reduction of packaging costs, open the path to vertical scalability and improve mechanical reliability.

Several solutions to combine MEMS and integrated circuits (IC) have been proposed in the literature and two major approaches can be identified: (1) system-on-chip (SoC) solutions, where mechanical elements and electronics are monolithically integrated on the same wafer; (2) multiple chips (multi-chip) solutions, where mechanical elements are manufactured on dedicated dies which are bonded to ICs [13].

1.2 Objectives

The main objective of this work is to experimentally demonstrate the possibility to integrate a magnetic field sensing system using a multi-chips solution based on standard industrial processes and to achieve sensing specifications, which are required for navigation in consumer products. The choice of multi-chip approach is driven by its higher flexibility and lower development costs with a quite rapid system development cycle.

1.3 Research Contribution

The research path followed during this project consisted of study and development of a magnetic field sensing system for consumer applications starting from specifications. First, the work focused on system level modelling and on the design of MEMS magnetometers trying to provide design criteria for consumer products. Basing on the available technology, a novel design approach for MEMS magnetometers insensitive to ambient acoustic noise is investigated and characterized, showing that the specifications required for consumer applications can be fulfilled with a very compact device by using standard industrial processes. Second, research activity focused on the design of both driving and readout electronics keeping an eye at a system level to guarantee resolution, power consumption, linear range required by the application. The resulting prototype is one of the first to show the feasibility of a Z-axis Lorentz force-based magnetometer implemented with a double chips solution and using standard processes: one chip hosts the mechanical element and in the second chip readout electronics is integrated.