Current Projects

Fingertip Implantable MEMS Tactile Sensors

In the United States, there are more than five million people are paralyzed due to spinal cord injury, stroke and other causes. In recent years, there are studies on brain-machine interface technology where the patients can use their mind to control robotic arms or their own paralyzed hands so that restoring some daily activities like grasping would become possible for these patients. In order to fully restore the functionality, fingertip sensors for feedback are required to work with these actuators. We are developing a fingertip implantable tactile system for normal and shear force measurement with wireless power and signal transmission circuit encapsulated. The package is made of fused silica material so that it satisfies the hermeticity and biocompatibility requirements for long term implantation purpose. CO2 laser assisted simultaneous fusion bonding and dicing of fused silica wafers has been successfully demonstrated for encapsulation of CMOS circuits. Additional work to further evolve this packaging technology to a normal and shear force sensor is under investigation.


  1. L. Du and M.G. Allen, “Silica Hermetic Packages Based on Laser Patterning and Localized Fusion Bonding,” 2018 IEEE 31st International Conference on Micro Electro Mechanical Systems (MEMS), Belfast, UK, Jan. 2018. (PDF)
  2. L. Du and M. G. Allen, “CMOS Compatible Hermetic Packages Based on Localized Fusion Bonding of Fused Silica,” Journal of Microelectromechanical Systems, vol. 28, no. 4, pp. 656-665, Aug. 2019. (PDF)

High Frequency Step Up Power Supply

The growing demand for miniaturized power devices is evident in numerous applications, from military and space to consumer electronics. As the size of the components in these power devices decreases, the allowable frequency increases. However, the use of higher frequencies introduces a new set of problems related to the switching components and passive components used in power supplies. These new challenges can be resolved at both the device level and the system level. At the device level, eddy current loss is significant in magnetic devices, especially at high frequencies. Because of this issue, there has been growing interest in using inductors with laminated cores that have high saturation flux densities. The layered cores reduce eddy current loss in order to allow for more efficient high frequency operation. At the system level, traditional converter designs work well for large inductors at low frequencies, but the same cannot be said for the newer designs. We are investigating new circuit topologies that will fully take advantage of the high frequency performance of the laminated magnetic cores. By using an efficient, miniaturized power supply, we will be able to significantly reduce the size of the overall electronics while still supplying a constant, very high or very low voltage.

Rapid, point-of-care diagnostic measurement of cytokines in vivo would advance patient care. A desirable approach to this measurement would be a transdermal or transvascular sensor fabricated in the shape of a needle, in which the sensing portion is formed on the needle tip and extends into the body medium of interest upon insertion. When contemplating the sensing approach for such a sensor, it is important to maintain a high degree of specificity to the protein of interest. Electronic-based sensing is a promising approach for femtomolar level specific detection of proteins due to its low cost, ease of miniaturization, and label-free operation. We’re developing a wafer-level microfabricated, needle-shaped impedance sensor for rapid, label-free detection of cytokines and other biomarkers. The sensor, consisting of a 20 μm x 20 μm micro-well array comprising 25 individual 2 μm-diameter wells embedded in a sensing tip, is lithographically configured on a laser micromachined fused silica needle. Label-free specific detection is achieved via functionalizing the microwells with antibody and monitoring the impedance change across the sensor electrodes due to the binding of target protein to the antibody.

Conference paper:

  1. N. Song, P. Xie, H. Oh, W. Shen, M. Javanmard and M. G. Allen, ” Wafer-Level Micromachining of Insertable, Label-free Cytokine Sensing Platforms”, BMES Annual Meeting, Philadelphia, USA, Oct. 2019.
  2. Song, P. Xie, M. Javanmard and M. G. Allen, “Microwell-array on a Flexible needle: a Transcutaneous Insertable Impedance Sensor for Label-Free Cytokine Detection”, MEMS 2018, Belfast, UK, Jan. 2018. (PDF).

Rechargeable, or secondary batteries, are required to power almost all of today’s increasingly complex technology. In most current commercial systems, high charge and discharge rates can lead to increased kinetic resistances and reduced energy density, while increasing size and active material mass can increase energy storage but reduce power. To bridge this application gap, MEMS techniques have been identified as promising methods to build high surface area battery electrodes that can be optimized to reduce these kinetic resistances while maintaining energy storage capability. To create batteries with potential to be manufactured on a large scale, these electrodes must be deterministically engineered and cost effective while providing this maximized performance. We are developing sequential, multi-layer electroplating as well as lithography and laser micromachining techniques to design these electrodes for use in a wide variety of applications. Additionally, new materials such as conductive and carbonized polymers are being investigated as both sacrificial and active layers to make the device fabrication more environmentally friendly and decrease costs, hopefully leading to enhanced commercial viability.    

The rapid development of small electronic devices puts forward high requirements for the components responsible for delivering power to these systems. Most power conversion circuits require magnetic components (inductors and transformers) to operate efficiently. However, these components have not been widely miniaturized. We are developing sequential, multi-layer electroplating as well as lithography and nano-lamination techniques to design the magnetic cores for small-scale transformers.

Human brains are bottom-up synthesized machines that utilize highly complex networks of simple computational units i.e. neurons for cognitive processing. Artificial neural networks, which borrow aspects of this network topology, have already successfully demonstrated human-level classification performances in image and speech recognition tasks. However, implementing these networks on existing computer architectures with markedly different topologies leads to increased power consumption and sub-optimal operating speed. Recent efforts have evolved specialized transistor-based “neuromorphic” circuits that sacrifice the scalability of bottom-up fabrication technologies for the determinism of their top-down counterparts. In contrast, we are designing all-passive, transistor-free circuits that are amenable to the inherently stochastic but scalable bottom-up fabrication approaches. We are investigating the performance of these circuits on benchmark datasets and we are exploring simple fabrication techniques to realize them. The results of our research, we believe, will establish new methods for designing systems that embody intelligence at the physical level.