I have a broad interest in anything small: micro-systems, nanotechnology, MEMS, microsensors and actuators for physical and bio-chemical parameters, and their integration with biological components at the same scale.

My research interests are:

Below is a portfolio of some recent research activities.

Zhao, C., Wood, G.S., Xie, J.B., Chang, H. Pu, S.H. and Kraft, M. A three degree-of-freedom weakly coupled resonator sensor with enhanced stiffness sensitivity. IEEE/ASME J. of MEMS, Vol. 25, Iss. 1, pp. 38-51, 2015.

Zhao, C., Wood, G.S., Xie, J., Chang, H. Pu, S.H. and Kraft, M. A force sensor based on three weakly coupled resonators with ultrahigh sensitivity. Sensors and Actuators A, Vol 232, pp. 151-162, 2015. doi:10.1016/j.sna.2015.05.011

These publications describe a recent, very exciting new trend in micro-sensor technology. By coupling several micro-resonators together, mode localization can be used as a transduction method. It uses the vibrational amplitude ratios of the resonators as an output metric. When the system is symmetric the ratio is unity, but it exhibits a very strong dependency on any symmetry breaking perturbation. The perturbation can be in form of a change in stiffness of one resonator - this is described in the publications, where the stiffness is modulated by an electrostatic force. We experimentally demonstrated an increase of sensitivity of approximately 23000 (!) compared to a single resonator of similar dimension for which a change in resonant frequency is observed. The chip comprises three coupled resonators fabricated in an SOI technology.

SEM of the chip consisting of three resonators coupled to its each other by electrostatic springs. The output metric is the ratio of the vibrational amplitudes of the left and right resonators in the presence of a pertubation. Block diagram showing the three resonators coupled to each. The coupling springs are realised as electrostatic springs; the other are silicon beams. Measured normalized sensitivity for various coupling stiffnesses using the vibrational amplitude ratio as output signal. The sensitivity can be up to 23000 higher when compared to using frequency shift as an output metric.

Mross, S., Zimmermann, T., Winkin, N., Kraft, M. and Vogt, H. Integrated multi-sensor system for parallel in-situ monitoring of cell nutrients, metabolites, cell density and pH in biotechnological processes. Sensors and Actuators, B, in press. doi:10.1016/j.snb.2016.03.086

Mross, S., Fürst, P., Pierrat, S., Zimmermann, T., Vogt, H. and Kraft, M. Enzyme sensor with Polydimethylsiloxane membrane and CMOS potentiostat for wide-range glucose measurements. IEEE Sensors Journal, Vol. 15, No. 12, pp. 7096–7104, 2015. doi: 10.1109/JSEN.2015.2470111

These publications describe a microchip for monitoring the parameters inside a bioreactor. Bioreactors are high tech machines that are used to produce (not only beer and wine but) vaccines, artificial tissue, antibodies, stem cells and many other biological compounds. Is is of paramount importance to monitor the parameters inside such bioreactors to ensure the quality of the final product and maintain high efficiency of the production process. The most important parameters are concentration of nutrients (e.g. glucose) and metabolites (e.g. lactate), pH and cell density. Currently, there are no integrated sensors solutions to measure these parameters. We have therefore developed a CMOS compatible microchip that can measure these parameters concurrently and in real time. It relies on amperometric enzyme sensors for glucose and lactate at high concentrations typically found in bioreactors, a cell density sensor using interdigitated fingers for impedance spectroscopy and a pH sensor based on an electrolyte-insulator-semiconductor structure.

Photograph of the multi-sensor chip with the sensing areas labeled. The size is 7.2mm x 7.2mm. Fabrication of the multi-sensor chip:
1. start with p-doped silicon, 2. definition of sensor area, 3. thermal oxidation and boron implantation, 4. Ta2O5 deposition, 5. deposition of lift-off resist, structuring and metal evaporation, 6. lift-off.
Cell density of a yeast cell culture and pH value in the solution. Good agreement with reference measurements were achieved. Lactate measurements in a culture
of Lactobacillus acidophilus. The red dots are spectroscopic reference measurements. 

A linear accelerator for levitated micro-objects
I. Sari and M. Kraft
Sensors & Actuators A, Vol 222, pp. 15-23, 2015

In this work the design, microfabrication, and characterization steps of a contactless linear accelerator is presented. The proposed design in principle can levitate conductive micro objects and accelerate or move them over a predefined trajectory - akin to a MEMS railgun. The levitation is realized using electromagnetic induction generated by a changing AC field whereas the propulsion is achieved through electrostatic forces from a controlled DC source. This is the first time in literature that such a hybrid design is used to accomplish this idea. It has been experimentally shown that the proposed design can levitate 1mm x 1mm sized and 7um thick micro objects to a maximum height of 75um and propel them forward continuously at a maximum average forward velocity of 3.6mm/s.

Optical image of a prototype chip. On the left track a micro-object made of 7um thick Aluminum can be seen. Close up of one track. To minimize contact forces dimples were fabricated on the electrostatic propulsion electrodes. Optical profilometer scan of the levitated Al plate (blue). The uniform colour indicates a constant levitation height parallel to the surface and the flatness of the Al-plate.  See above video of the MEMS railgun in action. Look at the right-most track.

Encapsulation of implantable integrated MEMS pressure sensors using polyimide-epoxy composite and Atomic Layer Deposition
P. Gembaczka, M. Goertz, Y. Celik, A. Jupe, M. Stühlmeyer, A. Goehlich, H. Vogt, W. Mokwa and M. Kraft
J. of Sensors and Sensor Systems. Vol 3, pp. 335-347, 2014

A novel approach to encapsulation techniques of an integrated pressure sensor suitable for implantation in the human body was investigated. If future implants should really benefit of the advantages and potential of microtechnology a new encapsulations techniques need to be developed, as bulky and rigid encapsulation based on titanium or ceramics currently used for implants negate the advantages of microtechnology. We show that Atomic Layer Deposition and a polyimide-epoxy composite is a promising approach to hermetically seal a microchip.

SEM of the pressure sensor chip encapsulated with polyimide-epoxy composite and tantalum pentoxide deposited by ALD. Optical image of the pressure sensor chip showing the polyimide-epoxy composite flowing around the pressure sensor membrane. Optical image of the entire pressure sensor module with the CMOS electronics under the polyimide-epoxy composite.

Two-dimensional ion trap lattice on a microchip for quantum simulation
Sterling, R.C., Rattanasonti, H., Weidt, S., Lake, K. Srinivasan, P., Webster, S.C., Kraft, M. and Hensinger, W.K.
Nature Communications, Vol. 5, Article number: 3637, 2014

This work reports the operation of a two-dimensional ion trap lattice integrated in a microchip capable of implementing quantum simulations of two-dimensional spin lattices. Our device provides a scalable microfabricated architecture for trapping such ion lattices with coupling strengths between neighbouring ions sufficient to provide a powerful platform for the implementation of quantum simulations. In order to realize this device we developed a specialist fabrication process that allows for the application of very large voltages. We fabricated a chip containing an array of microtraps forming a closely spaced ion lattice. We demonstrate reliable trapping of a lattice of ytterbium ions, measure long ion lifetimes and show the lattice is suitable for performing quantum simulations.
This was a collaboration between Prof. Hensinger's group at Sussex University who did the physics, and my group who did the chip design and fabrication.

SEM of the microfabricated chip with a honeycomb structure. Each honeycomb can trap one ion with electric forces (Paul trap). The cross-section of the chip, fabricated on an SOI wafer with 10um thick oxide. Note the unusual etch profile of the oxide. The ion chip loaded with ytterbium ions. Each white dot is exactly one trapped ion. The ions can be shuffled from one position to another and interact. Here, three interacting ytterbium ions are shown.

Design and implementation of an optimized double closed-loop control system for MEMS vibratory gyroscope
Chen, F., Yuan, W., Chang, H., Yuan, G., Xie, J. and Kraft, M.
IEEE Sensors Journal, Vol. 14, Iss. 1, pp. 184 – 196, 2014

Parameter optimization for a high-order band-pass continuous-time sigma-delta modulator MEMS gyroscope using a genetic algorithm approach
Chen, F., Chang, H., Yuan, W., Wilcock, R. and Kraft, M.
J. Micromech. Microeng., Vol.22,  No. 10, 105006, 2012


These publications describe the design and implementation of a advanced control system for a MEMS gyroscope. A high-order, band-pass electromechanical sigma-delta modulator was used for the sense mode of the gyroscope. We could demonstrate that the bias stability (measured by the Allan variance) was significantly improved. The electromechanical sigma-delta modulator was designed and optimized with a genetic algorithm implemented in Matlab/Simulink.
This was a collaboration between Prof. Honglong Chang's group at the Northwestern Polytechnical University in Xi'an, China, who fabricated the gyroscope sensing element, and my group who did the control system design.

The electromechanical sigma-delta modulator control system for the MEMS gyroscope. Due its complexity and nonlinear characteristics, it is impossible to optimize it analytically. Therefore, we used a genetic algorithm in Matlab/Simulink. The gyro sensing element fabrication at the Northwestern Polytechnical University in Xi'an, China in an SOI process. The gyro sensing element and the circuit implementation of the control system mounted on a rate table, ready for testing. The Allan Variance is a measure of the bias stability, the smaller the better. The red curve is with the sigma-delta control system and the green curve is open-loop.

Design, modelling, and evaluation of thermoelectric generators with hot-wire chemical vapor deposited polysilicon as thermoelement material
de Leon, M.T., Tarazona, A. Chong, H. and Kraft, M.
Journal of Electronic Materials, Vol. 43, Iss. 11, pp. 4070-4081, 2014

Parameter optimization for a high-order band-pass continuous-time sigma-delta modulator MEMS gyroscope using a genetic algorithm approach
de Leon, M.T., Tarazona, A. Chong, H. and Kraft, M.
J. Micromech. Microeng. Vol. 24, 085011, 2014


In this work we investigated the idea of using solar concentrators together with a new design of a thermoelectric generator (TEG). The TEG consisted of a round membrane with thermoelectric elements radially arranged on its periphery. Focusing light through a lens onto the membrane can generate a local hot spot, thus producing a large temperature gradient over the thermoelectric elements. This can significantly increase the power output of a thermoelectric generator. It similar to using a magnifying glass to burn a hole in a piece of paper.

The conceptual design of a solar thermoelectric generator (STEG). A lens is used to focus light on a round membrane and generate a large temperature drop across the thermoelectric elements. Optical image of the fabricated chip. A SOI wafer was used with etched trenches and refilled oxide. As thermoelectric materials simply Aluminum and Silicon were used. COMSOL simulation of the temperature distribution for 1W of solar input power. Simulated and measured data of the output versus input power.

Design and analysis of an SOI MEMS voltage step up converter
Gleeson, R. Kraft, M. and White, N. and Kraft, M.
J. Micromech. Microeng. Vol. 23, 114017, 2013

Modelling and analysis of a MEMS approach to DC voltage step-up conversion
Haas, C. and Kraft, M.
J. Micromech. Microeng. Vol. 14,  No. 9. pp. S114-S122, 2004


This work investigates the possibility of a MEMS DC-DC voltage step-up converter. The principle is very simple: charge up a MEMS capacitor to a voltage, isolate the capacitor, move on electrode away from the other, and since the charge cannot flow anywhere, the voltage has to increase. The approach is an alternative to charge pumps and is intended to interface with energy harvesters producing a low or varying output voltage that needs to be stepped-up to do something useful. Out work showed that the principle is feasible, but further work needs to be done to make it work in practice.

Conceptual sketch of a DC-DC voltage step-up converter. The force to move the electrodes apart comes from ambient vibration, hence the large proof mass. Simulation results showing how the voltage is increased when the electrodes of the MEMS capacitor are moved apart (in a cyclic manner). Prototype MEMS DC-DC voltage step-up converter. This is a version in which the electrodes are actively pulled apart by a MEMS actuator. Measurement results proofing the general validity of the approach. Due to leakage currents the charge is not well isolated and hence the chip did not work as well as intended.

Characterization of a mechanical motion amplifier applied to a MEMS accelerometer
Zeimpekis, I., Sari, I. and Kraft, M.
IEEE/ASME J. of MEMS, Vol. 21, Iss. 5, pp. 1032-1042, 2012

Many MEMS sensors rely on the detection of a very small displacement of a proof mass or membrane. The electronic interface inevitably contaminates the signal with noise. The idea behind this research was to pre-amplify the displacement with a mechanical lever. This has the advantage that a mechanical amplifier does not actively generate noise (in contrast to an electronic amplifier), therefore the signal to noise ratio can be increased and a sensor with a lower noise floor can be realized. We demonstrated this concept for an accelerometer.

Chip design of the accelerometer with the mechanical amplifier and SEM images of the capacitive comb fingers, located at the proof mass and at the output of the levers. Measurement on these two comb fingers allowed to characterize the mechanical amplification. SEM image of the chip prototype showing one quarter. Optically measured frequency response of the displacement of the comb fingers on the proof mass (lower curve) and at the output of the lever. The shape is the same, but there is a displacement gain of about 30dB. Video of the mechanical amplifier in action.

A dicing free SOI process for MEMS devices
Sari, I., Zeimpekis, I. and Kraft, M.
J. Microelectronic Engineering, Vol. 95, pp. 121-129, 2012

This work describes the development of a MEMS SOI process having the advantage that no dicing is required. DRIE is used to structure the device area, but also to etch spatially offset trenches into the top and bottom layers, which are used to separate the chips.  Furthermore, the handle wafer is removed underneath the device area. This is allows the fabrication of large underetched areas, for example for very high sensitive inertial sensors. The process has a very high yield (>95%) and required only two masks.

The process relies on two DRIE step and a HF vapour phase etch for release. Separation of the chips is achieved by etching not dicing. The handle wafer is removed underneath the device area. A processed 6-inch wafer with about 150 chips. The yield was >95%.