A New Method to Make Programmable Transducers in Microelectromechanical Systems (P-MEMS)

Lead Research Organisation: University of Bath
Department Name: Electronic and Electrical Engineering


Modern technologies have vastly benefitted from the miniaturised transducers developed in Micro-Electromechanical Systems (MEMS). Magnetic microdevices are one class of MEMS that demonstrate a significant potential for future applications in microrobotics, microfluidics, lab-on-chip, etc. For example, magnetic microfluidic chips can drastically reduce the costs and increase the throughput of DNA sequencing in genomic explorations and single-cell array analysis for cancer diagnosis. However, despite their great potential, the integration of magnetic material in microfabrication processes remains costly, time-inefficient and is therefore still an open research topic. For example, microtransducers manufactured by etching deposited layers have lower performance than bulk magnets of the same size. Even the most promising solutions for integrated magnetic materials in microfabrication processes, unfortunately deliver identical magnetic properties for the entire set of microdevices on a chip and cannot be individually tuned afterwards.

This project will develop a comprehensive solution to these problems. It will be delivered in three work packages (WP).

In WP1, this project will develop a new microfabrication process to realise on-chip programmable transducers in MEMS (p-MEMS). This process will integrate an array of magnetic microdevices with individual electrothermal microheaters. In this innovative technique the temperature of each microdevice increases in response to the applied electrical power to its corresponding heater. Applying an external magnetic field then produces the selective magnetic annealing. Therefore, exposing the entire magnetic microdevices on a chip to an external magnetic field and connecting selected heaters to electrical power will result in permanent magnetic changes only in the selected microdevices. Hence, this technique can develop various magnetic polarity patterns on a single chip by applying different combinations of external magnetic fields and selected heaters. WP1 will be carried out in close partnership with experts from BAE Systems who are particularly interested in the future micro and nano technologies.

In WP2, this research work will develop a new comprehensive micromagnetism model (M-MAG) to understand the magnetic behaviour of these microtransducers. The magnetic behaviour of thick microfabricated ferromagnetic (FM) layers in MEMS is different from thin films and bulk magnets. Thick layers are different from bulk magnets due to their atomic ordering after deposition. They are also different from thin deposited layers whose thicknesses of a few atoms impose certain constraints and assumptions that are not necessarily valid for thick layers. The new M-MAG model will be used to develop a computer aided design (CAD) tool for integrating the magnetic behaviour of microdevices into the available three dimensional micromechanical modelling tools. This will provide the multiphysics finite element analysis for the design of future microtransducers in p-MEMS.

In WP3, a prototype microfluidic chip will be developed using the p-MEMS process to test and verify the reliability of the process as well as the accuracy of the M-MAG model and simulations in the new CAD tool. There is a wide variety of applications for p-MEMS. Focusing on microfluidic applications will extend the cross disciplinary benefits of the proposed technique beyond Engineering and Physics. Just as programmable electronic integrated circuits enabled a wider community of non-expert users, the proposed research on p-MEMS will lead to a broader usage of these transducers emerging among other disciplines such as Biotechnology and Chemistry. Hence, WP3 will apply feedback from various end-users including experts from iGenomix UK.

The p-MEMS design kit including the layer thicknesses, material properties and layout design rules as well as the M-MAG model and the CAD tool will all be made freely available on the project webs

Planned Impact

Integration of ferromagnetic material in Microelectromechanical Systems (MEMS) will make significant impacts on the development of new microtransducers for future technologies. In response to major existing technological barriers, this project will deliver a comprehensive solution including (i) a microfabrication process for programmable transducers in MEMS (p-MEMS), (ii) a micromagnetic model (M-MAG) to develop a new CAD tool, and (iii) a microfluidic prototype chip with programming capabilities. This project will advance the knowledge of microfabrication and micromagnetism and develop new capabilities and skills among experts and non-experts. Achieving the objectives of this project will have direct and indirect impacts on the following main beneficiaries:

The p-MEMS design kit including microfabricated layer geometries, material properties and process steps required for developing new transducers in p-MEMS, will be made available on the project website. This will provide a low-cost and reliable platform for the MEMS community to design and manufacture new microtransducers, e.g. sensors and actuators. Just as in various Multi-user MEMS Processes (MUMP), this will give MEMS designers have free access to design kits including the process and material property database files as well as geometric design rules for developing new microdevices in a CAD environment.

The M-MAG model developed for this process will answer several questions such as magnetisation at elevated temperatures for thick deposited layers, which is still an open research topic in the Physics community. This model will be used to develop the new CAD tool for the simulation of the thermomagnetic and micromechanical behaviour of magnetised microdevices.

The proposed programming capabilities in this research have diverse end-user applications such as tagging and surveillance in harsh environments, controllable on-chip inductors in CMOS, microfluidic devices in lab-on-chip etc. This project will develop a prototype microfluidic device with on-chip reconfigurable microchannels in p-MEMS. These devices are in high demand by chemists, pharmaceutical researchers, biotechnologists, etc. as a high-precision enabling technology. The new microfluidic device will include an array of thermomagnetically programmable microvalves on the chip that connect microchannels to each other. The proposed prototype will have the capability of reconfiguring the links between microchannels in a very short time. Further to this, end-users with no expertise in magnetism or microfabrication will be able to programme various desired configurations on the microfluidic chip by simply connecting electrical inputs to the chip microheaters while the chip is exposed to an external magnetic field. Hence this project will bring the end-users closer to the advanced technologies.

The proposed research will make long-term impacts in growing the economy, healthcare and well-being of society.
The market value of MEMS products is expected to reach at US$ 66bn by 2021 (See reference 17 in the Case for Support). This huge market will lead to a high competition among manufacturers for low-cost microfabrication processes. This project will investigate the batch compatibility of every process step in p-MEMS to reduce the manufacturing costs. In addition, the programming capability will provide low-cost reconfigurable solutions for end-users such as biotechnologists as compared with ordering custom-made microchips.

In addition, high-throughput diagnostic technology, such as next-generation DNA sequencing, can help in finding the mechanism of diseases. The programming capability of the prototype microfluidic device is one step forward towards developing smarter diagnostic methods for individualised therapeutic treatment modalities.


10 25 50
Description We have investigated the deposition of different permanent magnetic films. The thin film deposition has already been investigated by various research groups and industries in nanometre ranges. However, the focus of this research is on thicker layers that could be used as mechanical actuators/sensors.
Therefore we have focused on electrodeposition.
The electrodeposition of CoPt has been successfully implemented in our cleanrooms. This included the deposition and photolithography of Copper electrodes etched on Silicon wafers.
In addition we have successfully implemented platinum microheaters deposited and etched on Silicon wafer, which need to be released from the thick Si substrate.
The next step is merging these two microfabrication processes to get one step closer to the final process.

We noticed the thickness of electroplated CoPt layers on SiO2 can reach only a few micrometres, which is not yet close to existing thick layers deposited on Si Tens of micrometres). The layers start cracking and delaminating as soon as the thickness reaches a few micrometre. We tried annealing the coper electrodes and different thicknesses deposited on different adhesion layers such as Cr. The marginal improvements are observed though still far from the existing thick layers deposited on Si.
We have also tried different atomic deposition of Cu and Au electrodes including Ebeam and Sputtering. Higher adhesion was achieved sing sputtering machine.
Exploitation Route The electrodeposition of hard magnets in micro-scale is a rare expertise we have developed in this project. This microfabrication expertise could be used by many other microfabrication rpocess developers. Although it is too early to comment on the benefits of our approach but we have stored the characterisation data corresponding to the manufactured magnetic samples.

We investigated the electrodeposition of thick CoPt films on Silicon wafers. The process starts with making Copper electrodes on the wafer. One potential approach to make these electrodes is the lift-off process. Although very well-known to the microfabrication community we had challenges in creating high adhesion between electrodes and substrate to the extent that the deposited electrodes peeled off during electrodeposition. We solved this problem by reducing the thickness of Chrome layer underneath as well as the Copper layer. Further to this the
Sectors Digital/Communication/Information Technologies (including Software),Electronics,Healthcare,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology

Description Alumni Fund
Amount £6,500 (GBP)
Funding ID GE-EE0105  
Organisation University of Bath 
Sector Academic/University
Country United Kingdom
Start 01/2020 
End 01/2021
Title Characterisation of developed microsamples 
Description The electrodeposited ferromagnetic samples has gone through different characterisation steps including Vibrating Sample Magnetometry and XRD. The extracted data is collated on the share drive of the project, which is stored at the university servers. 
Type Of Material Database/Collection of data 
Year Produced 2019 
Provided To Others? No  
Impact Although at the initial steps, this data base will be used to further characterise the whole programmable microtransducers in future. 
Description Deep Reactive Ion Etching (DRIE) at University College London 
Organisation University College London
Department Faculty of Mathematical and Physical Sciences (MAPS)
Country United Kingdom 
Sector Academic/University 
PI Contribution In our project, we made platinum microheaters (Less than 1micron thick) on Silicon substrates. In order to reach high temperatures the thick Si substrate (more than 300microns) underneath the heater must be removed. Otherwise it appears as a heat sink and avoids the high temperatures. In our initial design we intended to use wet etch to make deep trenches under heaters, which is available in our in-house cleanroom. However, the experiments show that isotropic wet etch creates wide shallow holes at the back side of the chip, where as narrow deep holes were desired for this project. DRIE is a well-known microfabrication process to etch thick substrates and release micro-resonators. However, this is only available at large cleanrooms and industrial microfabrication foundries. Oxford Instruments Plasma Technology agreed to do this for us but in a very late schedule. Academic colleagues from University College London also agreed to provide this service. We are now preparing test samples to calibrate the process for our substrate wafers.
Collaborator Contribution Colleagues form the Faculty of Maths and Physics at UCL has provided technical advice and access to facilities to run back etch process. This will allow our microheaters to be released from the substrate.
Impact We will inform the council about the outcomes in near future.
Start Year 2020
Description Electrodeposition of hard ferromagnetic material 
Organisation University of Bristol
Department School of Physics
Country United Kingdom 
Sector Academic/University 
PI Contribution We have successfully developed electrodeposited ferromagnetic (CoPt) samples on Silicon substrate. These samples will be integrated in our proposed programmable MEMS process technology. The magnetic properties of these samples must be extracted by high precision measurement instrument, which is not available to the University of Bath.
Collaborator Contribution Dr Chris Bell from the Department of Physics at University of Bristol, has kindly helped us to run characterisation steps on the magnetic layers deposited on Silicon substrates. Vibrating Sample Magnetometer is an expensive test instrument, which is conventionally used to extract the M-H curve for ferromagnetic material. Dr Bell has trained the postdoc researcher employed on this project to run initial experiments on as deposited samples. We have now annealed the ferromagnetic samples in our tube furnaces at the University of Bath. We will run further magnetisation characterisation steps to observe the annealing effect on the samples. This characterisation experiment is also essential to observe and characterise the effect of on-chip programming technique.
Impact There is not a published outcome yet. The measurement results has shown the magnetic properties of deposited layers are as expected. Over the next step we will repeat this experiment with annealed samples.
Start Year 2019
Description Sacrificial Layer Deposition 
Organisation BAE Systems
Country United Kingdom 
Sector Academic/University 
PI Contribution We will develop a microfabrication process wherein the magnetisation of ferromagnetic layers will be modified as a function of temperature raised by the heating layer underneath. This process consists of several steps: - The electrodeposition of CoPt on Silicon wafers has progressed successfully with some help from the School of Chemistry under supervision of Dr Adam Squires. - We have also observed some successful etching of silicon heaters on SOI (Silicon on Insulator) wafers. - A back-etch process is needed to create a deep trench isolation between the substrate and silicon heater. We have also etch handle wafer using KOH in our clean room at the University of Bath. The next step will consist of combining the steps above to ensure that heater can be createde underneath ferromagnetic layer. However prior to that we need to create a thin air gap between the
Collaborator Contribution Deposition of sacrificial layer is needed to create an air gap between heater (Silicon) and ferromagnetic layer (CoPt). This process step is not available in our clean rooms at the University of Bath. Fortunately this has been foreseen in the project description. Microfabrication experts from BAE Systems (the project industry partner) have advised about availability of this process at their microfabrication facilities at Filton, Bristol. Dr Ian Stureland and his colleagues have provided technical advice on this process step. We have been discussing this over the last few weeks. We are now ready to send the samples to BAE Systems clean rooms for the deposition of sacrificial layer.
Impact We expect the successful deposition of sacrificial layer at least in a prototype chip level. Then we will optimise this process step with further iterations and midifications to the thickness of the sacrificial layer.
Start Year 2019