Some content of this application is unavailable at the moment.
If this situation persist, please contact us atFeedback&Contact
Note: Text based on automatic Optical Character Recognition processes. Please use the PDF version for legal matters



The present invention relates to magnetic sensors, and more particularly to a high-sensitivity and ultra-low power consumption magnetic sensor comprising a magnetoelectric composite.


Magnetic field sensing provides critical information about battlespace threats. A wide variety of threats in the naval environment including ships, submarines, unmanned underwater vehicles (UUVs), and submerged or buried mines typically have magnetic signatures that can be identified at distance. UUVs employ electric motors that emit time-varying magnetic fields that can be expected to show unique and detectable signatures deriving from the currents and rotating magnetic assemblies of the motor. Sensitive and accurate determination of magnetic field strength combined with minimal power consumption would enable the targeting of autonomous undersea vehicles (AUVs) and other UUVs and extend mission duration/range by providing a system requiring minimal battery recharging.

There are several other technologies for magnetic field sensing, each of which is optimized for a particular magnetic field sensing range and has a number of disadvantages. Other than SQUID magnetometers, which require cryogenic temperatures for operation, sensors that operate at room temperature have a low maximum sensitivity on the order of several PT/VHZ. Additionally, the current state of the art for sensors operating in a similar range as said invention all require operational power of several mW or greater to drive a current (i.e. fluxgate, giant magnetoresistance or giant magnetoimpedance sensors). Magnetoelectric sensors that use some of the same basic principles as said invention are also very large in scale (Wang et a\., Adv. Mater. 2011, 23, AW 1), or utilize ferromagnetic resonance for detection (which is also a very high power method) (Nan et al, Sci. Rep. 2013, 3, 1985).

Despite advancements in Navy-specific technical research areas with relevance to magnetic sensors, present generation magnetic field sensors do not exhibit desired levels of sensitivity, <10"10 Tesla/Hz1/2, and power dissipation, <10 mW, in addition to low to zero drift due to local geomagnetic field variations.

In response to this challenge new types of materials (i.e. magnetoelectric (ME) composites) were developed and demonstrated to be very suitable for novel ME magnetic sensors. Currently these macro-ME laminates (ME composites) are most promising materials for breakthrough improvements.

ME materials couple the magnetic and electric properties to make a new type of sensor material that is sensitive to a wide range of applied magnetic field (l O-13 to 10"1 T) and exhibit an optimal frequency range of 10"3 to 102 Hz working in a passive mode (low power consumption). Laminated ME composite structures (typically containing intimately bonded ferroelectric (FE) and ferromagnetic phases (FM) phases) have higher efficiency than single-phase structures, and their maximum ME efficiency occurs at structural resonance. The ME sensors are superior to other existing technologies, with extraordinarily low power dissipation (~ 0.1 mW), compared with ~ 10- 100 mW for optically-pumped or fluxgate sensors. ME laminate sensors are lightweight and are suitable for miniaturization. Further improvement in signal to noise ratio and power consumption is expected with miniaturization of ME structures.

The use of thin films and reduced dimensionality with silicon based fabrication enables on-chip integration with traditional microfabrication techniques. The use of a piezoelectric drive and self-generated output signal allow for extremely low power operation compared to the current state of the art magnetic sensing technology.

The magnetoelectric magnetic sensors exploit newly demonstrated artificially engineered multiferroic (ferromagnetic + ferroelectric) structures that convert magnetic field strength directly to an electric voltage with high sensitivity (1 pT/VHz at low frequency) and with near-zero power dissipation.


The present invention provides a magnetoelectric (ME) composite comprising magnetostrictive and piezoelectric layers as the key components of a high-sensitivity and ultra-low power consumption magnetic sensor. It exploits the magnetically driven resonance shift of a free-standing magnetoelectric micro-beam resonator.

There are several advantages with the present invention. Integrated ME MEMS-scale magnetic sensor with sensitivity in 100 pT /Hz 1/2 range is comparable to the state in the art of magnetic sensor technology (Fluxgate etc.). Exploitation of thin film deposition techniques enhances the strain coupling between the two layers in ME sensor to achieve higher sensitivity than larger ME devices. Higher sensitivity is achieved by utilizing the doubly clamped geometry of the ME resonator as opposed to a single side clamped cantilever allowing broader range tuning. High efficiency and very low power consumption two orders of magnitude lower than in the state of the art magnetic sensor technology.

These and other features and advantages of the invention, as well as the invention itself, will become better understood by reference to the following detailed description, appended claims, and accompanying drawings.


FIG. 1A is a microscope image of patterned microbeam heterostructures before substrate release. FIG. IB is a schematic showing active layers and electrical contacts in a microbeam sensor.

FIG. 2 shows the frequency shift of a magnetostrictive microbeam resonator when placed in a magnetic field (from Kiser et al, Applied Physics Letters, 104 (7) 072408. (2014)).

FIG. 3 is a schematic of the silicon nitride scaffold fabrication process.


The present invention provides a passive, integrated MEMS based ME magnetic resonator as a magnetic sensor. Enhanced ME sensor performance is achieved through the use of a structural resonance MEMS approach combined with low-power modulation (mechanical or electric) techniques using the non-linear regime of the sensor to reduce the influence of background magnetic and vibrational noise. Resonant operation also allows for signal processing at frequencies above the Iff crossover for additional noise reduction (allowing for detection of DC and low-frequency near -DC magnetic fields). Noise reduction strategies make use of resonant frequency and low-cost wafer-scale batch fabrication of micro-nanoscale devices combined with ensemble averaging.

This invention realizes the first on-chip micro-resonator magnetic sensor with micron and sub-micron dimensionality (FIGS. 1 A and IB). The micro-beam is a bilayer thin film heterostructure of a magnetostrictive layer (Fe, Co, Ni, FeCo, FeCoV, FeGa ... ) strain couple to a piezoelectric layer (A1N, PMN-PT, BTO ... ). The magnetostrictive layer changes length under an applied magnetic field which alters the resonance frequency of the device. The heterostructure that makes up the micro-beam consists of an active magnetostrictive layer, which is strain coupled to a sensing piezoelectric layer, both grown atop a thin lattice matched metal seed layer (enabling high quality crystalline epitaxy and a conducting rear electrode). During operation the structure is excited to resonance using a low power piezoelectric crystal (PZT, PMN-PT ... ) or directly by the A1N piezoelectric layer.

When exposed to a magnetic field the active magnetostrictive layer responds causing a considerable resonance shift as a function of a magnetic field intensity. This approach has previously been demonstrated in clamped-clamped ME bimorphs with Metglas and FeNi (Finkel

et al, Appl. Phys. Lett. 98, 092905 (201 1)) as well as in ME cantilevers, where the shift in frequency /is related to the delta-E effect (Gojdka et al., Appl. Phys. Lett. 99, 223502 (201 1)).

Double-clamped stress-reconfigurable ME resonators can sense near DC magnetic field by determining the magnetic field induced shift in the resonant frequency resulting from a change in stress owing to magnetostriction (Kiser et al., Applied Physics Letters, 104 (7) 072408. (2014) and Staruch et al., Applied Physics Letters, 107 (3) 032909. (2015)).

The theoretical limit of detection of the proposed stress-reconfigurable sensor can be much improved compared to a one end fixed cantilever design (which has been developed but at much larger scales (Jahns et al, US Patent Application 20130289913)) due to higher Q (lower losses) and larger sensitivity proportional to frequency shift (df/dH) and reduced equivalent magnetic noise due to operation at higher resonance frequency and better resonant ME coupling and at the induced voltage. This resonance is sensed by the piezoelectric under-layer, which converts the planar strain to a voltage signal.

By sending the output of the piezoelectric layer to external circuitry, such as amplitude modulation or a phase locked loop, very low frequency magnetic fields can be detected at frequencies much less than the resonant frequency of the device.

The heterostructure is patterned into a microbeam using a series of patterning steps coupling with thin film deposition, ion milling and wet etching. The beams are then released by removing the silicon substrate using a high temperature KOH etch.

The invention also includes fabrication methods necessary to overcome extra-high internal materials stresses under sub-micron geometries which cause device self-destruction. To prevent such stresses from breaking the micro-beam, a fabrication procedure was developed using a 16 step process. Notable method inventions include epitaxial film growth optimization, a KOH patterned back etch using SiN as an etch stop, suspended reactive ion etching to avoid device contact and vertical processing baths to prevent hydraulic bending and stiction.

In one embodiment, a fabrication process is based upon building highly stressed multi-layered films on a removable silicon nitride scaffold, the silicon nitride can be partially or completely removed to form free standing double clamped cantilevers. Access to the silicon nitride is gained by back side etching completely through the silicon host wafer by KOH etching which has high selectivity between silicon and silicon nitride. A dry RIE such as a CF4 plasma, or other suable RIE technique is used to remove the silicon nitride scaffold from the back side. The fabrication process also includes initiation layers of the magnetostrictive film and the piezoelectric film utilizing a metallic buffer layer to establish an initial crystalline structure or adhesion layer conducive to achieving good magnetostrictive and piezoelectric quality thin films. The fabrication process includes the steps described below.

Fabrication Process utilizing Silicon Nitride Scaffold may include a number or all of the following steps (FIG. 3):

1. Deposit low-stress thermal silicon nitride (0-100 MPa tensile stress) on both sides of uniform double side polish Si wafers. The low-stress thermal silicon nitride can be low- stress low pressure chemical vapor deposition (LPCVD) silicon nitride, low-stress chemical vapor deposition (CVD) silicon nitride, low-stress plasma enhanced chemical vapor deposition (PECVD) silicon nitride, or low-stress atomic layer deposition (ALD) silicon nitride.

2. Deposit a contact layer serving as a seed layer (e.g., Pt, Hf, Ta, or any combination

thereof) on top side of wafer (serves as an initiation layer for the piezoelectric film and as a back side electrode).

3. Deposit/Sputter Al N or other piezoelectric layer at high temperature (450 °C or higher, preferably between 450 and 550 °C).

4. Partem photoresist / lift-off layer to define beam geometry.

5. Sputter metallic buffer layer for initiation/adhesion layer.

6. Sputter FeCo or other magnetostrictive film optimized for magnetostrictive property and stress. This process may be completed at elevated temperatures to perform a post- deposition anneal without breaking vacuum.

7. Lift off magnetostrictive film utilizing wet chemistry.

8. Photoresist process for piezoelectric layer etch to access under lying contact layer (Pt in this case).

9. Chemical Etch A1N or other piezo-layer to define first electrode.

10. Photoresist process for top side second electrode.

11. Electrode Metallization (thermal, ebeam or sputtering).

12. Chemical lift off process for first electrode definition.

13. Perform photoresist patterning of expanded openings back side aligned to the multi-layer beams.

14. RIE etch openings in the back side silicon nitride exposing Si windows.

15. Form a smaller opening at the front side of the wafer, centered under the beams. This can be done with a KOH etch on the back side - etch follows crystalline angles.

Alternatively, windows can be opened from the top with a CF4 plasma.

16. Back side RIE of the silicon nitride scaffold to release the multi-layer beams.

A second Fabrication process utilizes top side release of the beam by opening windows in the underlying Pt layer exposing the silicon and isotropically etching the silicon utilizing XeF2 sufficient to etch completely under the multi-layer film beams.

This ME integrated resonant sensor devices with on-chip silicon-based electronics to provide signal processing and communications functions in a compact module of micron-scale as integrated ME sensor elements with low power consumption and integrated tuning and readout mechanisms. Specialized low noise electronics can be developed for demonstrating equivalent magnetic noise reduction at resonance utilizing modulation (mechanical and/or magnetic) techniques.

The above descriptions are those of the preferred embodiments of the invention. Various modifications and variations are possible in light of the above teachings without departing from the spirit and broader aspects of the invention. It is therefore to be understood that the claimed invention may be practiced otherwise than as specifically described. Any references to claim elements in the singular, for example, using the articles "a," "an," "the," or "said," is not to be construed as limiting the element to the singular.